Attitude Change:
Multiple Roles for Persuasion Variables
Richard E. Petty
Ohio State University
Duane T. Wegener
Yale University
OVERVIEW
The O.J. Simpson “trial of the century” in the mid-1990s captured the attention of the
American populace more than any other public spectacle since the kidnaping of the Lindberg baby in the 1920s.
A prominent football player and popular sportscaster was charged with a gruesome double homicide. The
attorneys for the prosecution and defense were of various races and genders. The evidence presented on each
side was at times amazingly simple, visual, and emotional, and at times was verbal, abstract, and probably
incomprehensible to jurors. The witnesses included individuals of diverse styles, demeanors, and credibility.
The jurors, the recipients of the messages from these various sources, were themselves a mixed group of people
of diverse backgrounds, beliefs, and personal experiences who had to sift through the trial material and arrive at a
decision as to whether the defendant had been proven guilty or not. The context in which all of this took place
was at times tense and sad, and at times filled with humor and positive feelings. Not surprisingly, no experiment
has ever captured the extraordinary complexity inherent in this situation, yet almost all of the variables present in
this trial (and many not present) have been examined in the social psychological literature on attitude formation
and change. This chapter provides an overview of research on these diverse variables and addresses the
processes by which these variables are thought to result in influence.
Although it has become a cliché to say that the attitude construct is the most indispensable concept in
contemporary social psychology, this statement appears as true today as when Allport (1935) initially wrote it in
the first Handbook of Social Psychology (Murchison, 1935; see Allport, 1935, and Fleming, 1967, for historical
reviews of the attitude concept). Attitudes remain important as we enter the 21st century because of the
fundamental role that individuals’ attitudes, both explicit and implicit (Greenwald & Banaji, 1995), play in the
critical choices people make regarding their own health and security as well as those of their families, friends,
and nations. From purchase decisions provoked by liking for a product to wars spurned by ethnic prejudices,
attitudes help to determine a wide variety of potentially consequential outcomes.
Before turning to the relevant studies, it is useful to address some definitional issues. The term attitude is
used to refer to a person’s overall evaluation of persons (including oneself), objects, and issues. Thus, one’s
attitude refers to how favorably or unfavorably or how positively or negatively in general one views some object
of judgment such as “the defendant,” “capital punishment,” or “ice-cream.” These global evaluations can vary in
a large number of ways in addition to their extremity such as whether they are based on emotions (e.g., “seeing
the defendant makes me anxious”), beliefs (e.g., “capital punishment does not deter crime”), or past experiences
and behaviors (e.g., “the last time I ate ice-cream, I had an unpleasant allergic reaction”; see Breckler, 1984;
Zanna & Rempel, 1988), and whether they are internally consistent (e.g., being associated largely with positive
feelings, attributes, and behaviors) or ambivalent (e.g., composed of a combination of positive and negative
attributes, e.g., Kaplan, 1972). Much research has assumed that evaluative processing of information is quite
natural, pervasive and fundamental (Markus & Zajonc, 1985; Osgood, Suci, & Tannenbaum, 1957) in large part
because of the adaptive and functional purposes of holding attitudes (see Breckler, Pratkanis, & Greenwald,
1989; Eagly & Chaiken, this volume; for discussions of the structure and function of attitudes). Jarvis and Petty
(1996) suggested, however, that just as there are situational factors that influence one’s motivation to evaluate
(e.g., the need to form an opinion about an upcoming decision), so too are there individual differences in this
need. They observed that some people were higher in their “need to evaluate” than others and these individuals
were more likely to engage in on-line evaluation of a variety of stimuli and were more likely to hold opinions on
a diverse set of social issues than were those low in this need (see also Petty & Jarvis, 1996). Nevertheless,
1
regardless of the overall level of one’s propensity to evaluate, it is still the case that nearly everybody forms
evaluations of most common stimuli in their environment at some point (e.g., Bargh, Chaiken, Raymond, &
Hymes, 1996; Fazio, 1995; see Petty, Wegener, & Fabrigar, 1997, for more detailed discussion).
Given the power of attitudes to determine many actions, it is not surprising that billions of dollars,
deutchmarks, and yen are spent annually in an attempt to influence and change peoples’ evaluations of various
objects, issues, and people. Attitude change simply means that a person’s evaluation is modified from one value
to another. Change is often assessed relative to the person’s initial attitude. Polarization occurs when people
move in the direction of their initial tendency (e.g., an initially favorable person becomes even more favorable)
and depolarization occurs when they move in the opposite direction. In persuasion settings, change is typically
examined with respect to the position advocated in a communication. These change outcomes fall on a
continuum anchored at one end by maximal acceptance (e.g., changing the maximal amount in the direction of
some target position even if the change exceeds the target) and at the other by maximal boomerang (changing the
maximal amount in the direction away from the advocated position). It is important to distinguish these attitude
change outcomes from the processes that produce them. Acceptance processes are those that tend to move people
in the targeted direction, whereas rejection or resistance processes are those that help people resist change and
perhaps even move them significantly away from the target view. For each acceptance process documented in
the literature, there is typically a corresponding resistance process (e.g., pro versus counterarguing; positive
versus negative affect).
A categorical distinction between attitude formation versus attitude change is not used because the
accumulated research suggests that it is more useful to regard attitudes that are changed as falling along a
continuum ranging from non-attitudes (see Converse, 1970) to strong attitudes (see Fazio, 1986). That is, the
factors involved in moving an individual with no attitude to adopt a position favorable toward an advocacy are
more similar to the factors involved in making a person with a weak but existent attitude become more favorable
toward the advocacy than to the factors involved in making the same change in an individual with a strong initial
attitude (see Petty, Wegener, Fabrigar, Priester, & Cacioppo, 1993). The non-attitude/strong attitude continuum
is indexed by features such as how accessible the attitude is (e.g., Fazio, 1995), how much knowledge an
individual has about the attitude object (e.g., Wood, Rhodes, & Biek, 1995), and other factors (see Petty &
Krosnick, 1995, for reviews of the dimensions of attitude strength). Although this chapter focuses on attitude
change, the processes of change that are described and the general principles of change are also applicable to
changing more specific evaluative judgments -- such as whether the defendant in a trial is guilty or not. In fact,
even though there are functional differences between evaluative and non-evaluative beliefs, many of the
principles that are outlined for attitude change are also relevant to changing non-evaluative judgments.
Therefore, although the thrust of our chapter is on processes of attitude change, the term ?belief change” would
often be equally applicable.
Bringing about change in attitudes (or beliefs) by presenting facts and information in a relatively
objective fashion can be called education, whereas bringing about change by slanting information and evidence is
often referred to as propaganda (Zimbardo, Ebbesen, & Maslach, 1977). The latter term was popular in the
earlier part of this century (e.g., Doob, 1935), but because of its pejorative connotations has been replaced by the
more neutral terms, persuasion and attitude change (see Petty & Cacioppo, 1981). In the following sections,
some historical foundations of contemporary work on attitude change are reviewed and, using the Elaboration
Likelihood Model (ELM; Petty & Cacioppo, 1986a, 1986b) as a guide, the high- and low-effort processes that
have been hypothesized to account for changes in attitudes are discussed. Then, using the multiple roles aspect
of the ELM, work is reviewed on the effects of specific source, message, recipient, and context variables on
attitude change.
Historical Foundations
The formal discussion of principles of attitude change can be traced to the ancient Greeks (see Aristotle’s
Rhetoric and Topics), but as McGuire (1969) documented, also flourished in Cicero’s Roman Republic and
during the Italian Renaissance (see Quintillian’s Institutio Oratoria). It was not until the current century,
however, that ideas about persuasion were linked to empirical observations. Early work ranged from content
analyses of political propaganda in which the authors speculated about the attitudinal and behavioral
consequences of propaganda messages (e.g., Lasswell, Casey, & Smith, 1935), to case studies of the ebb and flow
2
of public opinion in which the authors speculated about the causes of observed shifts in attitudes (e.g.,
Lazarsfeld, Berelson, & Gaudet, 1944).
In reviewing the history of research on communication and persuasion, Delia (1987) noted that in
contrast to the mostly correlational approaches adopted in sociology and political science, the social
psychological approach (e.g., Allport, 1924) was largely experimental. That is, features of some influencing
agent (e.g., a propaganda message) were manipulated and their effects on attitudes were observed. For example,
in the 1930s, the Motion Picture Research Council conducted a series of studies to examine the impact of
different movies on children’s knowledge, attitudes, and behavior (e.g., Peterson & Thurstone, 1933). The
results of these studies were rather complex and provided an early warning that attitude change effects would
depend on a host of individual and situational factors. In addition to these large scale field studies, researchers in
the 1930s also conducted individual experiments with a more limited set of independent and dependent variables
that have more in common with current research practices (e.g., Knower, 1935).
Although research on attitude change had a significant presence in the first half of the 20th century, the
pioneering efforts of Carl Hovland and the eminent investigators he assembled at Yale in the 1950s, were
instrumental in making the study of attitude change one of the central foci of social psychology. During World
War II, Hovland directed the mass communication program within the Army’s Information and Education
Division. The research that emanated from this effort (e.g., Hovland, Lumsdaine, & Sheffield, 1949) as well as
the investigations conducted when Hovland returned to Yale after the war, continue to be highly influential. In
many instances, the Yale group was the first to examine systematically the variables that continue to be of
interest today such as source credibility (Hovland, Janis, & Kelley, 1953), individual differences (Hovland &
Janis, 1959), attitude structure (Hovland & Rosenberg, 1960), message order effects (Hovland, 1957), egoinvolvement (Sherif & Hovland, 1961), and many others.
If social psychological research on attitude change in the 1950s was dominated by the work of the Yale
group, research in the 1960s was led by the consistency theorists -- especially Leon Festinger (1957) and his
students and colleagues (see Abelson, Aronson, McGuire, Newcomb, Rosenberg, & Tannenbaum, 1968). In
contrast to the Hovland group which explored a wide variety of external variables and explained the persuasion
outcomes of these variables with a diversity of learning-type formulations, the consistency theorists were more
focused conceptually -- emphasizing the internal tension that was thought to result when people engaged in
actions that were inconsistent with their beliefs, attitudes, and values. The approaches taken by the Hovland and
Festinger groups are still influential today (for a recent discussion of the divergent research styles of these
groups, see McGuire, 1995a).
Contemporary Conceptualizations
By the 1970s, attitude change research had become so abundant that the whole enterprise was threatened
with collapse due to an embarrassment of conflicting findings and theories. Reviewers of the literature noted that
there were literally thousands of empirical efforts, but little conceptual coherence (e.g, Fishbein & Ajzen, 1972;
Kiesler & Munson, 1975). Sherif (1977, p. 370) bluntly stated that there was “reigning confusion in the area”
and a “scanty yield in spite of a tremendously thriving output.” Fortunately, this state of affairs did not last long
as new multi-process frameworks developed that accounted for many of the apparently conflicting findings. The
two most popular of these models, the Elaboration Likelihood Model (ELM) and the Heuristic-Systematic Model
(HSM) originated in doctoral dissertations in the late 1970s (Chaiken, 1978; Petty, 1977), and were subsequently
expanded into full persuasion theories (Chaiken, 1987; Chaiken, Liberman, & Eagly, 1989; Petty & Cacioppo,
1981, 1986a, 1986b). These models placed greater focus on the moderation and mediation of attitude change
effects and explained how the same variable (e.g., source credibility, mood) could have different effects on
attitude change in different situations, and how a given variable could produce the same persuasion outcome by
different processes in different situations. A key idea in these new frameworks was that some processes of
attitude change required relatively high amounts of mental effort, whereas other processes of persuasion required
relatively little mental effort. Thus, Petty and Cacioppo (1981) reasoned that most of the major theories of
persuasion were not necessarily competitors or contradictory, but operated in different circumstances. Later in
this chapter, this notion is used to organize the major processes of persuasion. Although these models share
many notions, the ELM and HSM had somewhat different conceptual parents. The ELM’s high effort central
route was based on cognitive response theory (Greenwald, 1968; Brock, 1967) whereas the HSM’s view of
3
detailed (systematic) message processing, in some early treatments at least, emphasized “effort exerted in
comprehending message content, not effort exerted in cognitive responding or thinking about message content”
(Eagly, Wood, & Chaiken, 1978). Today the models have more similarities than differences and can generally
accommodate the same empirical results, though the explanatory language and sometimes the assumed mediating
processes vary (additional coverage of some differences between the models is presented later).
Before undertaking our review, it is useful to consider the generic mediational analysis of attitude change
that has explicitly or implicitly guided most persuasion research in the 20th century. Some independent variable
has an impact on one’s affect (emotions), cognitions, or behavior toward some object, and this in turn has an
impact on one’s attitude toward that object. For example, a television commercial might induce pleasant feelings
in connection with a political candidate. Associating these pleasant feelings with the candidate could produce
favorable attitudes toward the candidate (i.e., commercial --> positive mood --> favorable attitude). Sometimes
the hypothesized causal sequence is more complex. For example, in the dissonance paradigm, a individual is
enticed to act in a way that is inconsistent with his or her attitude (behavior), which induces an aversive state of
tension (affect), which in turn leads to bolstering thoughts (cognition), which finally results in an attitude that is
consistent with the behavior (Cooper & Fazio, 1984).
Almost every conceivable causal sequence of affect, cognition, and behavior has been proposed to
account for attitudes in at least some circumstances. For example, in their two stage model of message repetition,
Cacioppo and Petty (1980) argued that moderate repetition enhanced people’s ability to cognitively respond to
the message in a relatively objective manner (stage 1). However, as repetition becomes excessive (stage 2),
negative affect sets in (i.e., the message becomes obnoxious and irritating) and this negative affect biases
subsequent cognitive responses to the message. It is also possible for a variable to exert parallel effects on affect
and cognition. For example, in his analysis of fear appeals, Leventhal (1970) postulated that fear induced both
cognitive responses designed to protect oneself from danger (such as accepting the advocacy) and emotional
responses aimed at protecting oneself from aversive arousal (such as avoidance). Roselli, Skelly, and Mackie
(1995) argued that persuasion in response to cognitive arguments was mediated by cognitive responses to the
message, but persuasion in response to affective arguments was mediated by both cognitive and affective
responses (see also Zuwerink & Devine, 1996). The accumulated literature makes it clear that although the
affective, cognitive, and behavioral bases of attitudes can be independent (Zanna & Rempel, 1988), they are often
inextricably interlinked as postulated by the consistency theorists. That is, if one’s affect changes, one’s
cognitive responses and behavioral tendencies typically change as well (e.g., Rosenberg, 1960). To the extent
that this does not occur, people remain in an ambivalent state that can itself be aversive (Priester & Petty, 1996;
Thompson, Zanna, & Griffin, 1995).
Although much of the early work on attitude change dealt with relatively complex issues (such as
whether atomic submarines should or should not be built, Hovland & Mandell, 1957), tenets of contemporary
theories have also been tested using simpler objects such as consumer products (e.g., Petty, Cacioppo, &
Schumann, 1983; Snyder & DeBono, 1989). Thus, although many of the examples in this chapter use complex
issues as attitude objects, many of these same attitude change processes are also applicable to simpler attitude
objects.
PROCESSES OF ATTITUDE CHANGE
As noted previously, contemporary persuasion theorists recognize that different processes can lead to
attitude change in different circumstances. Some of these processes require diligent and effortful information
processing activity whereas others proceed with relatively little mental effort. In this section, the Elaboration
Likelihood Model of persuasion is described, and some prominent factors that determine whether people will
tend to exert high or low amounts of mental effort in a persuasion situation are reviewed (the HSM is discussed
subsequently). Next, the persuasion processes that tend to require relatively high amounts of mental effort are
described in more detail. Following this, the persuasion processes that tend to require relatively low amounts of
mental effort are described.
The Elaboration Likelihood Model of Persuasion
The Elaboration Likelihood Model of persuasion (ELM; Petty & Cacioppo, 1981; 1986a) is a theory
about the processes responsible for attitude change and the strength of the attitudes that result from those
processes. A key construct in the ELM is the elaboration likelihood continuum. This continuum is defined by
4
how motivated and able people are to assess the central merits of a person, issue or a position (i.e., the attitude
object). The more motivated and able people are to assess the central merits of the attitude object, the more
likely they are to effortfully scrutinize all available object-relevant information. Thus, when the elaboration
likelihood is high, people will assess object-relevant information in relation to knowledge that they already
possess, and arrive at a reasoned (though not necessarily unbiased) attitude that is well articulated and bolstered
by supporting information (central route). When the elaboration likelihood is low, however, then information
scrutiny is reduced and attitude change can result from a number of less resource demanding processes that do
not require effortful evaluation of the object-relevant information (peripheral route). Attitudes that are changed
by low effort processes are postulated to be weaker than attitudes that are changed the same extent by high effort
processes. The hypothesis of both high and low effort attitude change comes from recognizing that it is neither
adaptive nor possible for people to exert considerable mental effort in thinking about all of the messages and
attitude objects to which they are exposed. In order to function in life, people must sometimes act as "cognitive
misers" (Taylor, 1981), but at other times it is more adaptive to be generous with one’s cognitive resources.
The elaboration likelihood continuum incorporates both a quantitative and a qualitative distinction (see
Petty, 1996). That is, as one goes higher on the elaboration continuum, central route processes increase in
magnitude, and as one goes down the continuum, central route processes diminish in magnitude. This
quantitative variation suggests that at high levels of elaboration, people’s attitudes will be determined by their
effortful examination of all relevant information, but at lower levels of elaboration, attitudes can be determined,
for example, by less effortful (careful) examination of the same information, or effortful examination of less
information (e.g., the person critically examines just the first argument in a message, but not the remaining
arguments). In addition, however, the ELM incorporates a qualitative distinction. For example, consider a
person who is exposed to a message with ten arguments. The high elaboration (central route) processor would
tend to think about most or all of the information. If motivation or ability to think was reduced, the recipient
might think about each argument less carefully, or think about fewer arguments (quantitative difference).
However, the ELM proposes that other (peripheral) mechanisms that do not involve thought about the substantive
merits of the arguments could also influence attitudes when the elaboration likelihood is low. For example, a low
elaboration processor might simply count the arguments and reason that “if there are ten reasons to favor it, it
must be worthwhile” (see Petty & Cacioppo, 1984). Note that this process is qualitatively different from the
argument elaboration process in that this mechanism does not involve on-line consideration of the merits of the
arguments but instead involves reliance on a rule of thumb or heuristic that the person generates or retrieves from
memory (see also Chaiken, 1987). Other relatively low effort peripheral mechanisms that are capable of
producing attitude change without processing the substantive merits of the arguments include: classical
conditioning (Staats & Staats, 1958; Cacioppo, Marshall-Goodell, Tassinary, & Petty, 1992), identification with
the source of the message (Kelman, 1958), misattribution of affect to the message (Petty & Cacioppo, 1983;
Schwarz & Clore, 1983), and mere exposure effects (Bornstein, 1989; Zajonc, 1968).
Two ELM notions -- the “tradeoff” hypothesis, and the “multiple roles” hypothesis -- have been the
subject of some confusion in the literature (e.g., Eagly & Chaiken, 1993; Stiff, 1986). The first hypothesis
postulates a tradeoff between the impact of central and peripheral processes on judgments along the elaboration
likelihood continuum. That is, as the impact of central route processes on judgments increases, the impact of
peripheral route processes on judgments decreases. Note that the tradeoff is not in the occurrence of central and
peripheral processes, but in the impact of these processes on judgments. For example, the presence of one’s
friend might invoke the heuristic, “I agree with people I like” (Chaiken, 1980), but under high elaboration
conditions, this heuristic would be subjected to careful scrutiny just as the arguments in a message are subjected
to scrutiny (Petty & Cacioppo, 1986a; Petty, 1994). If the heuristic is found to lack merit as an argument for
supporting the advocated view, then it would have little impact on one’s summary judgment. On the other hand,
if the heuristic was deemed cogent, then it would. This scrutiny of the heuristic for merit would be less likely
under low elaboration conditions where the mere invocation of the heuristic would be sufficient for persuasion.
It is important to note that the ELM tradeoff hypothesis implies a number of things. First, at most points along
the continuum, central and peripheral processes would co-occur and jointly influence judgments (Petty, Kasmer,
Haugtvedt, & Cacioppo, 1987). Second, however, movement in either direction along the continuum would tend
to enhance the relative impact of one or the other process (e.g., effortful scrutiny for merit versus reliance on a
5
heuristic) on judgments.
It is important to note that changing the relative impact of one process over another on attitudes does not
imply that the impact of any given variable (e.g., source expertise, mood) on judgments must increase or decrease
as one moves along the continuum. This is because of the multiple roles hypothesis. In essence, the multiple
roles notion is that any given variable can influence attitudes by different processes at different points along the
elaboration continuum (Petty & Cacioppo, 1986a). Thus, for example, consider whether a manipulation of
“beautiful scenery” in an advertisement for a vacation location should increase or decrease in impact as the
elaboration likelihood is increased. If a person was not thinking about the ad very much (low elaboration
likelihood), then the beautiful scenery might have a positive impact simply due to its mere association with the
target location much as it might have a similar positive impact on evaluations of a new car that was located in the
scenery (peripheral route). However, as the elaboration likelihood is increased and the scenery is processed for
its merits with respect to the product, then the impact of the scenery on attitudes could be increased in the ad for
the vacation location due to its perceived relevance and merit (or have the same impact but for a different
reason), but show decreased impact in the ad for the car due to its perceived irrelevance for this product (central
route; see also the later section on multiple roles for persuasion variables).1
Determinants and Dimensions of Elaboration
In sum, the ELM notes that attitude change can vary in the extent to which it is based on mental effort.
Persuasion researchers have identified a number of ways to assess the extent to which persuasion is based on
effortful consideration of information. Perhaps the most popular procedure has been to vary the quality of the
arguments contained in a message and to gauge the extent of message processing by the size of the argument
quality effect on attitudes (e.g., Petty, Wells, & Brock, 1976). Greater argument quality effects suggest greater
argument scrutiny. Other procedures include assessment of the number and profile of issue relevant thoughts
generated (Petty, Ostrom, & Brock, 1981). High elaboration conditions are associated with more thoughts (e.g.,
Burnkrant & Howard, 1984) and thoughts that better reflect the quality of the arguments presented (e.g., Harkins
& Petty, 1981). Also, correlations between message-relevant thoughts and post-message attitudes tend to be
greater when argument scrutiny is high (e.g., Chaiken, 1980; Petty & Cacioppo, 1979b), and high message
elaboration can produce longer reading or exposure times than more cursory analyses (e.g., Mackie & Worth,
1989).
According to the ELM, in order for high effort processes to influence attitudes, people must be both
motivated to think (i.e., have the desire to exert a high level of mental effort) and have the ability to think (i.e.,
have the necessary skills and opportunity to engage in thought). There are many variables capable of affecting
the elaboration likelihood and thereby influencing whether attitude change is likely to occur by the high or low
effort processes that are described in more detail shortly. Some of these motivational and ability variables are
part of the persuasion situation, whereas others are part of the individual. Some variables affect mostly the
amount of information processing activity whereas others tend to influence the direction or valence of the
thinking.
Perhaps the most important variable influencing a person's motivation to think is the perceived personal
relevance or importance of the communication (see Petty & Cacioppo 1979b;1990; Johnson & Eagly, 1989).
When personal relevance is high, people are more influenced by the substantive arguments in a message and are
less impacted by peripheral processes (e.g., Petty, Cacioppo, & Goldman, 1981). Of course, variables other than
relevance can modify a person's motivation to think about a message. For example, people are more motivated to
scrutinize information when they believe that they are solely responsible for message evaluation (Petty, Harkins,
& Williams, 1980), when they are individually accountable (Tetlock, 1983), when they recently have been
deprived of control (Pittman, 1994), and when they expect to discuss the issue with a partner (Chaiken, 1980).
Increasing the number of message sources can enhance information processing activity (e.g., Harkins & Petty,
1981; Moore & Reardon, 1987), especially when the sources are viewed as providing independent assessments of
the issue (Harkins & Petty, 1987). Messages that are moderately inconsistent with an existing attitude schema
can enhance processing over schema-consistent messages presumably because the former could pose some threat
that needs to be understood or some incongruity that needs to be resolved (Cacioppo & Petty, 1979b). Other
incongruities can also increase information processing activity, such as when an expert source presents
surprisingly weak arguments (Maheshwaran & Chaiken, 1991), when the message does not present the
6
information in a form that was expected (Smith & Petty, 1996), and when people feel ambivalent rather than
certain about the issue (Maio, Bell, & Esses, 1996). In addition to factors associated with the persuasive
message, issue, or the persuasion context, there are individual differences in people's motivation to think about
persuasive communications. For example, people who enjoy thinking (i.e., those high in need for cognition;
Cacioppo & Petty, 1982) tend to form attitudes on the basis of the quality of the arguments in a message rather
than on peripheral cues (see Cacioppo, Petty, Feinstein, & Jarvis, 1996, for a review).
Among the important variables influencing a person's ability to process issue-relevant arguments is
message repetition. As noted previously, moderate message repetition provides more opportunities for argument
scrutiny (e.g., Cacioppo & Petty, 1979b; Gorn & Goldberg, 1980), which will prove beneficial for processing as
long as tedium is not induced (Cacioppo & Petty, 1989; Cox & Cox, 1988). External distractions (e.g., Petty et
al., 1976), fast presentations (Smith & Shaffer, 1995) external pacing of messages (such as those on radio or TV
rather than in print; Chaiken & Eagly, 1976; Wright, 1981), time pressures on processing (e.g., Kruglanski &
Freund, 1983), enhancing recipients’ physiological arousal via exercise (e.g., Sanbonmatsu & Kardes, 1988),
placing recipients in an uncomfortable posture (Petty, Wells, Heesacker, Brock, & Cacioppo, 1983), and
rendering the message difficult to understand (e.g., Ratneshwar & Chaiken, 1991) all decrease substantive
message processing and should increase the impact of peripheral processes. Interestingly, even though a number
of studies have examined differences in the actual ability of recipients to process a persuasion message, little
work has examined differences in perceived ability to process. Thus, a message that appears technical or overly
quantitative (Yalch & Elmore-Yalch, 1984) may reduce processing not because it interferes with actual ability,
but because it interferes with a person’s perceived ability to process (e.g., “it’s probably too complicated for me,
so why bother”).
Individual differences also exist in the ability of people to think about a persuasive communication. For
example, as general knowledge about a topic increases, people can become more able (and perhaps more
motivated) to think about issue-relevant information (Wood et al., 1995). Knowledge is only effective to the
extent that it is accessible, however (e.g., Brucks, Armstrong, & Goldberg, 1988). When knowledge is low or
inaccessible, people are more reliant on simple cues (e.g., Wood & Kallgren, 1988). To date, researchers have
not examined differences in perceived knowledge. It is reasonable that even if actual knowledge differences
were constant, people with less perceived expertise might be less likely to process a message and be more
influenced by peripheral processes.
Finally, it is important to note that in most communication settings, a confluence of factors determines
the nature of information processing rather than one variable acting in isolation. For example, ending arguments
with rhetorical questions rather than statements can increase thinking about a persuasive message if the questions
follow the arguments and motivation to think about the message would normally be low (Petty, Cacioppo, &
Heesacker, 1981). On the other hand, if people are already motivated to think about the message or the rhetorical
questions precede the message arguments, then the use of rhetoricals can actually disrupt the normal processing
that would have occurred (see Howard, 1990; Petty et al., 1981). Although the effects of many individual
variables on information processing have been examined, there is a shortage of studies examining the potential
interactions possible when multiple ability and motivational variables are combined.
Relatively Objective Versus Biased Information Processing
The variables already discussed, such as distraction or need for cognition, tend to influence information
processing activity in a relatively objective manner. That is, all else being equal, distraction tends to disrupt
whatever thoughts a person is thinking (Petty et al., 1976). The distraction per se does not specifically target one
type of thought (e.g., favorable or unfavorable) to impede. Similarly, individuals with high need for cognition are
more motivated to think in general than people low in need for cognition (Cacioppo, Petty, & Morris, 1983).
They are not more motivated to think certain kinds of thoughts over others. Some variables, however, are
selective in their effects on thinking. For example, people are facilitated in thinking favorable thoughts and
disrupted from thinking negative thoughts when they are instructed to move their heads in an up and down
(“yes”) manner. Moving one’s head in a side to side (“no”) manner, appears to facilitate negative thinking and
disrupt positive thinking (Wells & Petty, 1980; see also Förster & Strack, 1996).
Default biases. Before considering other variables, like head nodding, that induce specific biases in
information processing, it is useful to consider the default biases that can influence judgment. Specifically,
7
research has suggested that in the absence of contrary information, people tend to assume that what others say is
true (Gilbert, 1991), and that unfamiliar objects and people in our environment are good. The latter has been
called the positivity bias (Kanouse & Hanson, 1971; Peeters & Czapinski, 1990) or the positivity offset
(Cacioppo & Berntson, 1994). However, if a source is known to be a liar, and this information is retrieved,
statements from this source would not be assumed to be true even if no specific issue-relevant information to the
contrary was available. Also, even though unfamiliar objects and people are generally assumed to be good,
people would not necessarily assume that an unfamiliar insect was good even if no explicit negative information
about the new insect was available (since the category “insect” contains many negative instances whereas the
category person contains more positive instances). Interestingly, however, Cacioppo, Gardner, and Berntsen
(1997) have suggested that there is a default positive bias to evaluative processing. They demonstrated that when
people were presented with 6 completely neutral statements about an unfamiliar insect, judgments were more
positive than when this information was not available. This could be because of a default positive bias to
evaluative processing as suggested by Cacioppo et al., or it might result from application of Gricean rules of
conversation. Grice (1975) argued that people expect information from others to be informative and relevant. If
the most relevant information the experimenter can provide about an insect is neutral, one might safely assume
that there is little negative information to convey (e.g., the insect must not be poisonous, or surely this would
have been included). The Gricean interpretation would suggest that if the experimenter presented 6 neutral items
about an unfamiliar vacation location however, people might assume that if this is the most relevant information,
there must be little positive information to convey and attitudes could be less positive than in the absence of any
information.
Motivational and ability factors in bias. The ELM accommodates both relatively objective and relatively
biased information processing by pointing to the motivational and ability factors involved. Regarding
motivation, the ELM holds that motivation is relatively objective when no apriori judgment is preferred and a
person’s implicit or explicit goal is to seek the truth “wherever it might lead” (Petty & Cacioppo, 1986b). In
contrast, a motivated bias can occur whenever people implicitly or explicitly prefer one judgment over another.
This is similar to what Kruglanski (1990) has called a “need for specific closure.” A wide variety of motivations
can determine which particular judgment is preferred in any given situation. For example, if the reactance motive
(Brehm, 1966) is aroused, people will prefer to hold whatever judgment is forbidden. If balance motives (Heider,
1958) are operating, people would prefer to adopt the position of a liked source but distance themselves from a
disliked source. If impression management motives (Tedeschi, Schlenker, & Bonoma, 1971) are operating,
people would prefer to hold whatever position they think would be ingratiating. If self-affirmation motives
(Steele, 1988) are high, people would prefer the position that would make them feel best about themselves, and
so forth. Importantly, many of these biasing motives could have an impact on judgments by either the central or
the peripheral route. For example, invocation of reactance could lead to simple rejection of the forbidden
position without much thought or to rejection because of intense counterarguing of the position. Which occurs
(central or peripheral bias) would depend on other variables such as whether the person was motivated (e.g., high
personal relevance) or able (e.g., high distraction) to think.
The ELM holds that biased processing can occur even if no specific judgment is preferred (i.e., if based
on motivational factors alone, processing would be relatively objective). This is because ability factors can also
determine bias. For example, some people might simply possess a biased store of knowledge compared to other
people. If so, their ability to process the message objectively can be compromised. Recipients with a biased
store of knowledge typically will be better able to see the flaws in opposition arguments and the merits in their
own side compared to recipients with a more balanced store of knowledge (cf., Lord, Ross, & Lepper, 1979). In
addition, variables in the persuasion situation can bias retrieval of information even if what is stored is
completely balanced and no motivational biases are operating. For example, a positive mood can increase access
to positive material in memory (e.g., Bower, 1981). In general, biases in processing a persuasive message are
fostered when the message contains information that is ambiguous or mixed rather than clearly strong or weak
(Chaiken & Maheswaran, 1994).
Bias not only impacts judgments by influencing the nature of on-line information processing, but also can
influence judgments by biasing the integration of information. For example, even if people have an unbiased
store of information in memory, and a balanced set of elaborations is generated, certain ideas might be given
8
greater weight than others in forming a judgment (e.g., Anderson, 1981). In fact, considerable evidence exists for
the proposition that negative information is often given more weight than equally extreme positive information
(see Fiske, 1980; Kanouse & Hanson, 1971; but see Skowronski & Carlston, 1989). This might account for the
finding that in order to move one’s rating from the bad to the good side of a rating scale requires not 51%
positive attributes associated with an object, but over 60% (Lefebvre, 1985). Although, there are a number of
ways to account for this negativity bias in information integration (see Cacioppo & Berntsen, 1994; Peeters &
Czapinski, 1990), one appealing explanation is that the negativity bias stems from the default tendency that was
noted earlier for people to prefer and expect positive information. That is, if people expect positive information,
then negative information would seem especially diagnostic and would be weighted more heavily. In addition, if
positive information is expected and preferred, then negative information would be more surprising or
threatening either of which could enhance attention to the information and processing of it (Baker & Petty, 1994).
This suggests, however, that in situations in which negative information is expected or preferred, positive
information should be perceived as more diagnostic or processed more (Smith & Petty, 1996).
Bias correction. It is important to note that just because some motivational or ability factor results in
biased information processing, this does not mean that a biased judgment will result. This is because people
sometimes attempt to correct for factors they believe might have unduly biased their evaluations (e.g., Petty &
Wegener, 1993; Wilson & Brekke, 1994). Initial discussions of such phenomena described these corrections as
proceeding only in one direction (i.e., a ?partialling” process moved ?corrected” assessments away from reactions
activated by the biasing stimulus; e.g., Martin et al., 1990; Schwarz & Bless, 1992). More recent discussions
have noted (and empirically shown) that corrections can proceed in different directions depending on recipients’
theories of how the biasing event or stimulus (e.g., an attractive source) is likely to have influenced their views
(e.g., Wegener & Petty, 1995). Specifically, according to the Flexible Correction Model (see Wegener & Petty,
1997), in order for corrections to occur, people should: (a) be motivated and able to identify potentially biasing
factors, (b) possess or generate a naive theory about the magnitude and direction of the bias, and (c) be motivated
and able to make the theory-based correction. In some cases, integrative processing of the information (e.g.,
Schul & Burnstein, 1985) could make it difficult for people to correct for the biasing effect of an individual piece
of information that contributed to an overall evaluation. That is, even if motivated to correct, people might not
have the ability to do so. When people are motivated and able to correct, however, theory-based corrections can
result in reversals of typical persuasion effects. For example, in one study, when people were made aware of
possible biases due to source attractiveness, an overcorrection led an unattractive source to be more persuasive
than an attractive one (Petty & Wegener, 1995; see Wegener & Petty, 1997, for additional discussion of
correction processes in persuasion and other settings).
High Effort Processes
As outlined above, a number of factors influence the likelihood that people will allocate mental effort in
persuasion situations. However, when people engage in high effort processing, what do they do? Through the
years, a number of high effort processes have been proposed. Many of the theoretical positions discussed in this
section were originally proposed -- and some still hold them to be -- general models of attitude change (e.g., see
Fishbein & Middelstadt, 1995). As was noted previously, however, within the ELM, the impact of these high
effort processes on attitudes is viewed as falling along a continuum such that these processes have a larger impact
on judgments when ability and motivation to scrutinize the central merits of a position or object are relatively
high.
A variety of relatively high-effort processes have been discussed over the years. These include message
learning/reception processes, cognitive responses, probabilogical/expectancy-value processes, information
integration, and at least in its original formulation, dissonance-produced reassessments of attitude-relevant
thoughts and actions.
Message Learning/Reception Processes
Our starting point in discussing contemporary high-effort information processing approaches is the
Communication-Persuasion matrix model outlined by William McGuire in previous editions of this handbook
(1969, 1985). McGuire began as a valuable contributor to the Hovland group and as an important consistency
theorist (e.g., McGuire, 1960a). His matrix model and related work provided an essential bridge from the
Yale/Hovland work to more contemporary information processing models (e.g., McGuire, 1964).
9
McGuire’s analysis is generally compatible with the Hovland/Yale view that attitude change depends on
a series of information processing stages described as: attention, comprehension, learning, acceptance, and
retention of the message and its conclusion (Hovland et al., 1953). A person was assumed to engage in these
“output” or “mediating” steps to the extent that the persuasive source, message or context (“input” factors)
provided incentives (i.e., rewards) for doing so. McGuire’s (1968; 1989) model incorporates and significantly
elaborates these ideas. At the same time, McGuire proposed an elegant simplification of the output factors by
noting that persuasion is often most dependent upon factors related to the reception of message arguments and
yielding to them. Thus, some variables might decrease the likelihood of attitude change by making reception
more difficult (e.g., distraction from the message content by a secondary task), whereas others might decrease
persuasion by making yielding less likely (e.g., a poorly reasoned message). From this perspective, some of the
most interesting variables were those that might influence reception and yielding in opposite ways. For example,
as the self-esteem or intelligence of message recipients increased, reception might increase (i.e., because higher
levels of these characteristics would likely increase attention and comprehension) but yielding might decrease
(e.g., because these people would likely have greater confidence in their initial opinions or would be more critical
of new information; Eagly & Warren, 1976). For variables that have opposite effects on reception and yielding,
persuasion should be maximal at a moderate level of the variable -- the ?compensation principle” (McGuire,
1968). This perspective provided considerable flexibility over the original message learning approach in that
either reception or yielding could become more important in a given persuasion setting. For example, if a
message consisted of complex but compelling arguments, factors affecting reception might be most related to
persuasion outcomes (because most people would yield to compelling arguments), but if a message consisted of
simple but somewhat weak arguments, factors related to yielding might take on greater importance (see McGuire,
1968).2
Although some studies have examined the curvilinear hypothesis from this model (see Rhodes & Wood,
1992, for a review), most of the attention devoted to the reception/yielding perspective has examined the notion
that reception is a prerequisite to persuasion. In particular, research has focused on the question of whether an
audience needs to comprehend and remember the message arguments in order for persuasion to occur. Regarding
comprehension of the arguments, research clearly suggests that if a message is gibberish, persuasion can be
reduced when compared to a comprehensible message (Eagly, 1974), unless, perhaps, if the arguments in the
message were very weak, and processing the intact message would have produced boomerang.
Considerable attention has been paid to the issue of whether people must learn and remember the
message arguments in order for persuasion to occur. Consequently, researchers have examined the extent to
which argument learning (and retention), as assessed by message recall or recognition, is related to persuasion.
Perhaps surprisingly, early research provided little evidence for a link between message recall and either initial
persuasion or its persistence over time (see Eagly & Chaiken, 1993; McGuire, 1969, for reviews).
There are several possible reasons for low attitude change-argument memory relationships. Perhaps the
most important is that a pure reception/learning model does not take into account the idiosyncratic evaluations
individuals have of the arguments. That is, one person can find an argument very compelling and relevant,
another only slightly convincing, and a third might find the same argument to be completely ridiculous and
irrelevant (Petty, Ostrom, & Brock, 1981; see section on cognitive responses). Consistent with McGuire’s (1968)
suggestions, however, when recalled information is weighted by a person's idiosyncratic evaluation of it, attituderecall relationships are increased (e.g., Chattopadhyay & Alba, 1988).
In some situations, however, there might be a substantial relation between memory for message
arguments and attitude change. For example, if a judgment is to be made when no information has been
presented on the topic for some time, or if people were distracted from careful on-line evaluation of the
information when it was presented, then people are more likely to base attitudes or other judgments on the
implications of the information that can be recalled if that information is perceived relevant at the time of
judgment (Hastie & Park, 1986). However, because the messages in the persuasion studies showing recallattitude correlations (e.g., Haugtvedt & Petty, 1992; Mackie & Asuncion, 1990) contained reasonably strong
arguments, it is not clear if the memory-recall relationships that were observed were due to the fact that people
favorably evaluated the arguments they recalled at the time of judgment, or whether they simply made inferences
of validity based on the mere number of arguments that they could recall (Petty & Cacioppo, 1984). That is,
10
attitude-memory correlations can stem from either central or peripheral routes to persuasion.
It is somewhat ironic that although the learning/reception perspective was originally formulated in a way
that suggests a high-effort process, the circumstances in which message recall relate most closely to attitudes
might actually be those in which relatively little on-line evaluation of message arguments takes place (i.e.,
passive learning of arguments used as a cue or evaluated later rather than active on-line consideration of them).
Specifically current research suggests that the correlation between message recall and attitudes should be
relatively high if: (a) the likelihood of elaborating the message arguments at the time of message exposure is
relatively low, (b) an unexpected judgment is required following message exposure, and c) no simple cues (e.g.,
source expertise; strong prior attitude) are available or attended to by the recipient that could produce an attitude
at the time of judgment in the absence of on-line argument scrutiny. In such circumstances, in order to express an
opinion, people would likely attempt to recall the information presented and base their judgment on what they
can remember.
In sum, although the reception hypothesis has a long history, and some consider it to be proven
(Chaiken, Wood, & Eagly, 1996), our view is that accurate understanding or learning of message arguments is
not necessary for persuasion to occur. Rather, as emphasized by the cognitive response approach discussed
shortly, a person’s subjective perception of understanding the arguments is more important than actual
understanding or learning of the information presented. A person can completely misperceive the message
arguments (score zero on accurate reception) and nevertheless generate favorable responses to what was
inaccurately perceived. In addition, peripheral processes (e.g., classical conditioning) can influence attitudes in
the absence of reception of message arguments. Nevertheless, it is likely that the role of message reception is
probably underestimated in the typical persuasion study, in part because the variance in reception is often too low
to detect effects (Eagly & Chaiken, 1984; 1993).
Cognitive Responses
Thoughts in response to messages. In response to numerous findings of low attitude-recall correlations
(which challenged the message learning view implied by the Yale approach of the 1950s), researchers -especially at Ohio State University -- developed the cognitive response approach to persuasion (e.g., Brock, 1967;
Greenwald, 1968; Petty, Ostrom, & Brock, 1981). For example, Greenwald (1968) proposed that it was not the
specific arguments in a message that were associated with the message conclusion (or attitude object) in memory.
Rather, a person's idiosyncratic cognitive responses or reactions to the message arguments were paired with the
conclusion and were thus responsible for persuasion or resistance (see also, Kelman, 1953; McGuire, 1964). The
decay of these cognitive responses were also postulated to be responsible for the persistence of attitude change
(Love & Greenwald, 1978). Of course, at the most basic level, individuals can only cognitively respond to
something that they have received. In the cognitive response analysis, however, a person's thoughts can be in
response to incorrectly perceived arguments as well as correctly perceived arguments. In fact, the thoughts can
be to the message conclusion in the absence of receipt of any arguments. The original cognitive response
approach was also sufficiently general that it accommodated cognitive responses elicited by the tone of the
message or by extra-message factors such as the source or context of the message. In any case, to the extent that
a person's cognitive responses were favorable, persuasion was the postulated result, but to the extent that they
were unfavorable (e.g., counterarguments, source derogations, emotion-laden thoughts), resistance or even
boomerang was possible.
The cognitive response approach has generated a considerable body of evidence consistent with the view
that in certain situations people spontaneously produce thoughts during message presentation and the favorability
of these thoughts is a good predictor of post-message attitudes and beliefs (see reviews by Eagly & Chaiken,
1993; Petty & Cacioppo, 1986a). In a typical cognitive response study, message recipients list or verbally report
their thoughts either during or after the message. Studies in support of the cognitive response approach have
shown that: (a) physiological activity indicative of information processing (e.g, facial EMG) is elevated when
cognitive responding is presumed to occur (e.g., Cacioppo & Petty, 1979a); (b) thought profiles show the same
pattern as the attitude measure in response to some manipulation (e.g., the manipulation produces increased
persuasion and increased favorable thoughts and/or decreased unfavorable thoughts; e.g., Osterhouse & Brock,
1970); c) the polarity of these thoughts (e.g., positive minus negative thoughts) is a good predictor of the postmessage attitude (e.g., Mackie, 1987), and (d) removing the effect of some manipulation on thoughts eliminates
11
its effect on attitudes, whereas the reverse does not occur (e.g., Insko, Turnbull, & Yandell, 1974; see Cacioppo
& Petty, 1981b, for a review).
Although most studies simply categorize thoughts into the valence dimension of favorable, unfavorable,
and neutral categories, other coding schemes are possible. For example, thoughts can be based mostly on affect
or cognition (e.g., Ickes, Robertson, Tooke, & Teng, 1986) and can reiterate the message content or include selfgenerated material (e.g., Greenwald, 1968). It is also possible to distinguish between cognitive responses that are
based on argument scrutiny (e.g., counterarguments) versus thoughts that are based on simple cues (e.g., source
derogations). Consistent with the multi-process models, it is the former type of cognitive response that best
predicts attitudes when elaboration is high; the latter category of thoughts predicts attitudes best when
elaboration is low (e.g., Chaiken, 1980; Petty & Cacioppo, 1979b, 1984). Although considerable work has
addressed factors that determine the extent of thinking and the content of thinking about a persuasive message at
the time a message is presented, little work has addressed the recurrence of thinking in persuasion contexts. That
is, what factors determine whether people return to thinking about the message or issue at some point following a
communication (see Martin & Tesser, 1996; Petty, Jarvis, & Evans, 1996; for discussion).
Thoughts when no message is present. Thoughts do not only determine the direction and extent of
attitude change when a message is presented. The powerful and persisting effects of completely self-generated
messages were shown in early research on "role-playing" (e.g., Janis & King, 1954; Watts, 1967). A consistent
research finding is that active generation of a message is a successful strategy for producing attitude change (e.g.,
McGuire & McGuire, 1996). It doesn’t seem to matter whether people generate the message because they are
assigned a position to take, or they select a position based on a desire to communicate with another person. For
example, people tend to tell others what they want to hear (Tesser & Rosen, 1975), and construction and delivery
of such biased messages can produce attitude change in the transmitter (see Higgins, 1981). Change tends to be
greater when people generate messages regarding attributes that a target possesses than attributes that a target
lacks (McGuire & McGuire, 1996).
Furthermore, self-generated attitude changes tend to persist longer than changes based on passive
exposure to a communication (e.g., Elms, 1966). Attitude changes that come about due to active generation of
arguments might be more persistent to the extent that they are based on more extensive processing of attitude
relevant information than are attitude changes due to passive receipt of messages (and argument generation might
make the arguments more accessible than when they are passively received; Greenwald & Albert, 1968; Slameka
& Graf, 1978). Similarly, people who are asked to imagine hypothetical events come to believe that these events
have a higher likelihood of occurring than before thinking about them (e.g., Sherman, Cialdini, Schwartzman, &
Reynolds, 1985). Furthermore, self-generation of explanations has been shown to be a powerful way to establish
or change beliefs, and beliefs based on the generated explanations are remarkably impervious to change (e.g.,
Anderson, Lepper, & Ross, 1980; Sherman, Zehner, Johnson, & Hirt, 1983).
Tesser (1978) and his colleagues have examined the effects of merely asking someone to think about an
issue, object, or person. In one early study, Sadler and Tesser (1973) introduced research participants to a likable
or dislikable partner (via a tape-recording). Some participants were instructed to think about the partner whereas
others were distracted from doing so. The thinking manipulation polarized judgments of the partner. More
recent research shows that the polarization effect requires that people have a well-integrated and consistent
schema to guide processing and that they are motivated to utilize this issue-relevant knowledge (e.g., Chaiken &
Yates, 1985). In the absence of these conditions, such as when motivation to think is low or when the issuerelevant information in memory is not highly interconnected (and consistent), mere thought is associated with
attitude moderation (e.g., Judd & Lusk, 1984; see Tesser, Martin, & Mendolia, 1995, for a review of mere
thought research).
In a variation of this mere thought procedure, Wilson and colleagues have examined the effects of asking
people to think about why they hold the attitudes they do (e.g., why do you like strawberry ice-cream? see Wilson
et al., 1989). The primary conclusion from a series of studies is that evaluations following a reasons analysis
tend to be overinfluenced by cognitive factors and underinfluenced by affective factors. Thus, decisions
following an analysis of reasons can be maladaptive because people consider only a subset of the real reasons
they like or dislike something. In a similar vein, Levine, Halberstadt, and Goldstone (1996) demonstrated that
when asked to analyze why they liked a target object prior to making judgments of a series of previously
12
unfamiliar (and not very important) stimuli, people use different dimensions of judgment on different trials.
Thus, when analyzing reasons, judgments of objectively similar objects were rated as less similar than when
reasons were not analyzed prior to evaluating the objects.
Probabilogical and Expectancy/Value Processes
Although the probabilogical (e.g., McGuire, 1960b; Wyer, 1974) and expectancy/value (e.g., Fishbein &
Ajzen, 1975; Rosenberg, 1956) models have largely been discussed in relation to the underlying structure of
attitudes (see Eagly & Chaiken, this volume; Petty, Priester, & Wegener, 1994), each is also applicable to
thoughtful processes of attitude change. For example, the underlying structure of at least some attitudes might
consist of 2 premises that lead to a conclusion:
Premise 1:
Candidate A favors tax reduction.
Premise 2:
Tax reduction is good for the country.
Conclusion: Candidate A is good for the country.
Therefore, one could change the attitude toward Candidate A by changing either of the premises, either in terms
of the content of the premise (e.g., candidate A doesn’t really favor tax reduction) or in terms of the likelihood
that the premise is true (e.g., there is a .2 rather than a .9 probability that premise 1 is true; McGuire, 1960a,
1960b; Wyer, 1974). To the extent that the attitude has an extensive horizontal or vertical structure (McGuire,
1981), it might take change in a greater number of premises in order to effect noticeable change in the target
attitude. Such a view also indicates that one might find changes in attitudes that are not even mentioned in a
persuasive message if the attitudes are logically related to claims in the message. Consistent with this view,
research shows that a message about abortion attitudes can produce changes in people’s attitudes toward the
unmentioned issue of contraception (e.g., if abortion is viewed as especially bad, then avoiding abortion by using
contraception might be viewed as more desirable than would otherwise be the case; see Mugny & Perez, 1991;
Dillehay, Insko, & Smith, 1966). This ?indirect change” might even occur if the “direct change” of the targeted
attitude is not evidenced (e.g., if change of the targeted attitude is inhibited by social pressure such as a norm
against agreeing with a minority source, but no such pressure is associated with the indirect conclusion; Mugny
& Perez, 1991). In fact, simply responding to an attitude item on one issue can produce an attitude polarization
effect on a related issue (Henninger & Wyer, 1976; Judd, Drake, Downing, & Krosnick, 1991), perhaps because
of respondents perceiving links between the issues -- links that dictate changes in the related opinion as a means
of maintaining consistency among attitudes in the system.
Versions of the probabilogical model have addressed both the likelihood and desirability of premises of
syllogisms like that presented above (McGuire, 1960a; McGuire & McGuire, 1991; Wyer, 1973). That is,
researchers have examined the determinants of both the desirability and likelihood of various belief statements.
For example, consider Premise 1 above that “Candidate A favors tax reduction.” If this was an argument in a
message advocating Candidate A for president, it would be effective to the extent that people found tax reduction
to be desirable and they believed that it was true that Candidate A really favored tax reduction. A theme of this
chapter is that desirability (evaluative) judgments can be made based on effortful or non-effortful analyses, but it
is less commonly recognized that likelihood or truth judgments can also vary in the extent of effort required.
Some theorists, for example, have posited that likelihood judgments involve a relatively effortful consideration of
the potential antecedents of the event (McGuire & McGuire, 1991). Other theorists, however, have argued that
likelihood judgments can be based largely on perceptions of familiarity of the idea (Arkes, Boehm, & Xu, 1991;
Wyer, 1991). As noted earlier, research suggests that propositions are initially assumed to be true unless
contradictory information is retrieved which allows one to determine that the proposition is false (Gilbert, 1991).
To the extent that this formulation is valid, people should be biased toward assuming truth when the elaboration
likelihood is low (Gilbert et al., 1993). Similarly, as noted earlier, people might be biased toward assuming
goodness when the elaboration likelihood is low (Cacioppo & Berntsen, 1994; Peeters & Czapinski, 1990).
Repetition under low elaboration conditions appears to enhance these default biases. That is, when the
elaboration likelihood is low, repeated exposure to statements can make them seem more true (Arkes et al., 1991)
and repeated exposure to objects can make them seem more good (Zajonc, 1968). As the elaboration likelihood
increases, however, propositions and objects should be judged either true or false and good or bad, depending on
the outcome of one’s effortful processing.
Regardless of whether the likelihood and desirability of a proposition are determined effortfully or not,
13
these parameters need to be combined. The most common approach to combining likelihood and desirability
comes from expectancy-value theories which analyze attitudes by focusing on the extent to which people believe
that the attitude object is linked to important values or is associated with positive versus negative outcomes (e.g.,
Rosenberg, 1956; see Bagozzi, 1985 for review). In the most influential formulation, Fishbein and Ajzen (1975)
hold that the desirability of each attribute associated with an object is weighted by the likelihood that the object
possesses the attribute, and the products of the likelihood and desirability components are summed over all
attributes. The major implication of expectancy-value theories for attitude change is that a persuasive message
will be effective to the extent that it produces a change in either the likelihood or the desirability component of an
attribute that is linked to the attitude object (see Fishbein & Ajzen, 1981; Petty & Wegener, 1991, for
discussions). Given the wide applicability of the Fishbein and Ajzen model and the extensive number of studies
documenting the link between attitudes and the likelihood and desirability components of beliefs, it is surprising
that relatively little work on attitude change has been guided explicitly by this framework. Nevertheless, existing
research supports the view that messages can influence attitudes by changing either the desirability or the
likelihood component of beliefs (e.g., Lutz, 1975; MacKenzie, 1986). Although some advocates of this
perspective assume that virtually all attitude change takes place through relatively effortful processes for
changing likelihood or desirability components (e.g., Fishbein & Middlestadt, 1995), there are some indications
that these processes account for more variance in attitudes when elaboration likelihood is high (e.g., when
message recipients are high in need for cognition, Wegener, Petty, & Klein, 1994; or have high levels of topicrelevant knowledge, Lutz, 1977).
Information Integration
Regardless of whether one conceptualizes high elaboration processes as cognitive responses, assessments
of likelihood and desirability of attributes of the attitude object, or as learning and retention of attitude-relevant
information, these information units must be combined in some manner to form an overall attitudinal reaction.
Two combinatory rules have generated the bulk of research attention. Fishbein and Ajzen's expectancy-value
formulation described earlier provides one such model (i.e., summation of the likelihood X desirability products
for each unit of information). An alternative integration formula is provided by Anderson's (1971) information
integration theory. In contrast to the additive rule specified by Fishbein and Ajzen, Anderson posits that the
pieces of information in a communication (or one's cognitive responses; Anderson, 1981) are typically combined
by a weighted averaging process. Specifically, the person's evaluation of the salient information is weighted by
the importance of the information for the judgment and is averaged with the person's weighted initial attitude to
form a new attitude.
Perhaps the biggest strength of the information integration model is simultaneously its greatest weakness,
namely, the very flexible use of the weighting parameter (Eagly & Chaiken, 1984; Petty & Cacioppo, 1981).
That is, the fact that the weighting parameter can take on either positive or negative values of different
magnitudes across different situations gives the model a virtually unlimited domain of applicability.
Unfortunately, the lack of a priori prediction concerning when and why the weighting parameter for certain
variables changes magnitude or sign makes the model more useful as a descriptive rather than explanatory tool
for understanding attitude change.3
Cognitive Dissonance Theory
Of the various theories proposing a motive to maintain cognitive consistency (e.g., Heider, 1958; Osgood
& Tannenbaum, 1955), the most prominent is the theory of cognitive dissonance (Festinger, 1957, 1964). In
Festinger's original formulation of the theory, two elements in a cognitive system (e.g., a belief and an attitude;
an attitude and a behavior) were said to be consonant if one followed from the other. The elements were
dissonant if one belief implied the opposite of the other. Of course, two elements could also be irrelevant to each
other. Festinger proposed that the psychological state of dissonance was aversive and that people would be
motivated to engage in cognitive activity in order to reduce it. Because people in a dissonant state are motivated
to achieve a particular outcome, their effortful information processing activity is clearly biased (i.e., of two
equally plausible interpretations, the interpretation most consistent with the other salient cognitive elements is
clearly preferred). The most obvious solution to dissonance is to engage in cognitive work to modify one of the
dissonant elements (i.e., self-generated attitude change). Strategies other than changing the dissonant elements
are also possible, however. For example the person could try to generate cognitions that make the dissonant
14
elements consistent with each other (bolstering). Alternatively, the person could try to minimize the importance
of one of the dissonant cognitions (see Simon, Greenberg, & Brehm, 1995).
Perhaps the most studied dissonance situation involves inducing people to engage in some behavior that
is inconsistent with their attitudes (Brehm & Cohen, 1962). For example, one common way of producing
dissonance in the laboratory is by inducing a person to write an essay that is inconsistent with the person's
attitude under high choice conditions and with little incentive (e.g., Zanna & Cooper, 1974). Because behavior is
usually difficult to undo, dissonance can be reduced by changing beliefs and attitudes to bring them into line with
the behavior (i.e., convincing oneself that the behavior reflects one’s true position). In the most famous
dissonance experiment, students at Stanford University engaged in the quite boring task of turning pegs on a
board (Festinger & Carlsmith, 1959), and then were induced to tell a waiting subject that the task was interesting
for either a sufficient ($20) or an insufficient ($1) incentive. When later asked how interesting they actually
found the task to be, students reported that the task was more interesting the less they were paid. A large number
of studies have examined the novel reverse-incentive predictions of dissonance theory in a wide variety of
situations. For example, dissonance theorists have discovered that people will come to like a group more the
more unpleasant the initiation required to get into the group (Aronson & Mills, 1959). People were found to
report liking an exotic food more the more dislikable the person who induced them to try the food (Zimbardo,
Weisenberg, Firestone, & Levy, 1965). Also, it was found that people's evaluations of two objects was more
discrepant after a choice between them than before the choice took place. Following a choice, biased cognitive
activity renders the chosen option more favorable and the unchosen option less favorable (Brehm, 1966).
It is now clear that many of the situations described by Festinger as inducing dissonance produce the
physiological changes and perceptions of unpleasantness predicted by the theory. Evidence for this proposition
has been found both with self-report (Elliot & Devine, 1994) and physiological measures (e.g., Elkin & Leippe,
1986; Losch & Cacioppo, 1990). In addition, the dissonance effect has been shown to disappear when people can
misattribute their unpleasant feelings to some other source (e.g., Fried & Aronson, 1995; Zanna & Cooper,
1974), or people are tranquilized to reduce their presumed discomfort (e.g., see Fazio & Cooper, 1983). Recent
work has also suggested that the attitude change associated with dissonance is more likely undertaken in order to
avoid the negative affective experiences rather than the physiological ?tension” proposed by the theory (Elliot &
Devine, 1994; Losch & Cacioppo, 1990).
It is also important to note that the mere performance of an inconsistent action does not always produce
dissonance. One reason for this is that some people may not have a need for consistency (Cialdini, Trost, &
Newsom, 1995). Alternatively, some theorists have questioned Festinger's view that inconsistency per se
produces discomfort in people, and have suggested that it is necessary for people to believe that they have freely
chosen to bring about some foreseeable negative consequence for themselves or other people (e.g., Cooper &
Fazio, 1984). Thus, if telling a waiting research participant that a boring task is interesting (Festinger &
Carlsmith, 1959) results in no harmful consequence (e.g., because the waiting participant doesn't believe you),
there is no dissonance. But, if the harmful consequences are high (e.g., the waiting participant decides to stay
and take part in the experiment rather than study for an exam), dissonance occurs and attitude change toward the
task results (see Calder, Ross, & Insko, 1973). Other theorists argue that inconsistency is involved, but the
inconsistency must involve a critical aspect of oneself or a threat to one's positive self-concept (e.g., Aronson,
1969; Greenwald & Ronis, 1978).
Interestingly, theorists from both camps have argued that proattitudinal advocacy (making a speech that
is consistent with one's attitudes) can also produce dissonance under certain conditions. Advocates of the
negative consequences view argue that proattitudinal advocacy can induce dissonance if the proattitudinal
advocacy ends up having negative consequences (Scher and Cooper, 1989; but see Johnson, Kelly, & LeBlanc,
1995). Advocates of the self-inconsistency view also argue that proattitudinal advocacy can produce dissonance
if, as a result of the advocacy, people feel hypocritical (which threatens self-esteem; Stone, Aronson, Craine,
Winslow, & Fried, 1994).
Although research has supported both the negative consequences and the self-inconsistency predictions,
disentangling these viewpoints has proven difficult. The reason for this is that each framework generally can
accommodate the results generated by the other. The conceptual problem stems from the fact that freely
choosing to bring about negative consequences is clearly inconsistent with most people's views of themselves as
15
rational, caring individuals. That is, choosing to bring about negative consequences is inconsistent with one's
positive self view. However, it is also true that when people do something inconsistent with their positive selfviews, the resulting feeling of guilt, shame, stupidity, or hypocrisy is an aversive consequence (Aronson, 1969).
That is, by choosing to violate one's self view, one has freely chosen to bring about an aversive outcome (see
Petty, 1995).
A third viewpoint on the causes of dissonance is provided by self-affirmation theory (Steele, 1988).
According to this framework, dissonance is not produced by inconsistency per se, being responsible for negative
consequences, or the distress of self-inconsistency in particular, but rather stems from a violation of general selfintegrity. According to this viewpoint, actions produce dissonance only when the behavior threatens one's "moral
and adaptive adequacy" (see also Tesser & Cornell, 1991). The self-consistency and self-affirmation points of
view have much in common, but they differ in their predictions of whether high or low self-esteem people should
be more susceptible to dissonance effects. The self-consistency point of view argues that high self-esteem
individuals would experience the most dissonance by engaging in esteem threatening behavior because such
actions are most inconsistent with their favorable self-conceptions. The self-affirmation point of view suggests
that low self-esteem individuals should show stronger dissonance effects because high self-esteem individuals
can more easily restore self-integrity by thinking about the many positive traits they have. Unfortunately, the
research evidence on this question is mixed with some studies showing greater dissonance effects for low selfesteem individuals (Steele, Spencer, & Lynch, 1993), and other studies showing greater dissonance effects for
high self-esteem persons (Gerard, Blevans, & Malcolm, 1964).
In sum, it is not clear whether dissonance results from cognitive inconsistency per se, the production of
aversive consequences, inconsistency in specific aspects of the self-concept, threats to general self-integrity, or
some other process. In fact, each of these mechanisms might be responsible for dissonance, but in different
situations. One attempt at integration is Stone and Cooper’s (1996) proposal that dissonance results whenever
one’s behavior violates some self-standard (see Higgins, 1989). That is, Stone and Cooper argue that dissonance
can stem from failure to behave in a manner consistent with how a person thinks he or she wants to be (ideal-self)
or how the person thinks he or she should be (ought-self), or how others want you to be or think you should be
(i.e., normative standards). In this framework, depending on which self-standard is salient, any given act may or
may not induce dissonance. Among the most important implications of this integration are that the selfconsistency view of dissonance prevails whenever personal self-standards are salient and violated, but the
negative consequences view holds whenever normative standards are salient and violated. In addition this view
holds that self-affirmation will only reduce dissonance when the affirmation is on dimensions irrelevant to the
dissonant act. Relevant affirmations will serve to increase dissonance suggesting self-consistency motives
remain at the core of dissonance theory.
Regardless of the exact motivational underpinnings of dissonance, the evidence clearly indicates that
attitudinally discrepant actions can result in a reanalysis of the reasons why a person engaged in a certain
behavior (or made a certain choice), and cause a person to rethink the merits of an attitude object. The end result
of this effortful but generally biased cognitive activity can be a relatively enduring change in attitude toward the
object.
Low Elaboration Processes
If the high effort processes just outlined were the only ones that could produce attitude change, it would
appear that changes typically require a diligent message recipient -- one who is willing (at a minimum) to
passively learn the information presented, and might also assess the likelihood and desirability of the attributes of
the object, generate new information and implications of the information, and weight and combine the
information to form an overall judgment. If so, attitude change should become less likely as such processes are
impaired. According to contemporary multi-process perspectives, however, there are a variety of situations in
which attitudes are formed or changed without a great deal of effortful attention or consideration of the
substantive information. When this occurs, a number of “peripheral” mechanisms can be responsible for
changing attitudes. In fact, a number of theories that were first proposed as general theories of attitude change
were comprised of mechanisms that do not require much (if any) effortful scrutiny of the merits of attitude
objects. Our discussion of relatively low effort change processes begins with conditioning and priming
mechanisms that associate affect with the attitude object, and concludes with a consideration of simple cognitive
16
inference and other processes that can produce attitude change in the absence of learning or effortfully evaluating
the substantive information or arguments presented. Although these peripheral mechanisms vary in the extent of
mental effort they require, none requires the recipient to personally evaluate the relevant pieces of information
for their central merits.
Associating Affect with the Attitude Object
Classical conditioning. One of the most primitive means of changing attitudes involves the direct
association of "affect" with objects, issues, or people through classical conditioning. Considerable research has
shown that attitudes can be modified by pairing initially neutral objects with stimuli about which people already
feel positively or negatively. For example, people’s evaluations of words (e.g., Staats & Staats, 1958), other
people (e.g., Griffit, 1970), political slogans (e.g., Razran, 1940), products (e.g., Gresham & Shimp, 1985), and
persuasive communications (e.g., Rogers, 1983) have been modified by pairing them with affect producing
stimuli (e.g., unpleasant odors and temperatures, the onset and offset of electric shock, harsh sounds, pleasant
pictures, and elating versus depressing films; e.g., Gouaux, 1971; Staats, Staats & Crawford, 1962; Zanna,
Kiesler, & Pilkonis, 1970). Contractions of muscles associated with positive or negative experiences (e.g.,
smiling versus frowning; Strack, Martin, & Stepper, 1988; or flexing arm muscles associated with moving objects
toward rather than away from oneself; Cacioppo, Priester, & Berntson, 1993) can also influence evaluative
responses. That is, when an initially unconditioned (neutral) stimulus (UCS) is encountered along with a
conditioning stimulus that is already strongly associated with positive or negative experiences (CS), the initially
neutral stimulus can come to be associated with the positive or negative experiences (the conditioned response).
Consistent with the view that conditioning effects can be obtained by non-thoughtful means, DeHouwer,
Baeyens, and Eelen (1994) reported evidence of evaluative conditioning even when the unconditioned stimuli
were presented subliminally. Also consistent with the notion that conditioning processes largely act as a
peripheral means to establish or change attitudes, Cacioppo et al (1992) showed that classical conditioning using
electric shock had a greater impact on initially neutral non-words (which, of course, were not associated with any
preexisting meaning or knowledge) than on initially neutral words; see also Shimp, Stuart, & Engle, 1991).
Similarly, isometric flexion versus extension of upper arm muscles during evaluative processing (i.e., processes
associated with approach and withdrawal, respectively) has been shown to influence preferences for neutral nonwords more than for neutral words (Priester, Cacioppo, & Petty, 1996).
Affective Priming. In a procedure similar to conditioning, people are presented with affect-inducing
positive or negative material (e.g., pictures) just prior to receipt of the target stimulus. This "backward
conditioning" or "affective priming" procedure has proven successful in modifying attitudes. For example,
research participants who were exposed to subliminal positive photos (e.g., a group of smiling friends)
subsequently rated a target person performing normal everyday activities more positively than participants
exposed to negative photos (e.g., a bucket of snakes; Krosnick, Betz, Jussim, & Lynn, 1992). Murphy and
Zajonc (1993) found that the effectiveness of this type of affective priming procedure may be dependent on
presenting the primes outside of conscious awareness. That is, when positive and negative affective primes
(smiling and frowning faces) were presented just prior to a target stimulus (a Chinese ideograph), attitudes
toward the target were influenced when the primes were presented subliminally, but not when they were
presented visibly (see also Murphy, Monahan, & Zajonc, 1995). One possible reason for such an effect might be
that visibly presented priming stimuli are noticed as obviously irrelevant to perceptions of the targets. Therefore,
this “blatant priming” might instigate an intentional avoidance of the perceived effects of the emotional primes
(Martin et al., 1990; Petty & Wegener, 1993).
In sum, studies of affect association (i.e., classical conditioning and affect priming) show that primitive
affective processes are most likely to influence attitudes toward objects that have little meaning and for which
people possess little or no knowledge, and when opportunities for processing are low (see also Zajonc, this
volume). This does not mean, however, that affect will have an impact on attitudes only when the likelihood of
elaborating attitude-relevant information is low. As discussed in the section on multiple roles for persuasion
variables, affect can also modify attitudes when the elaboration likelihood is quite high. In such cases, however,
the processes that lead to those changes are different than those discussed here.
Inference-Based Approaches
Attribution theory. Although attributional processes can, themselves, vary in the amount of cognitive
17
effort required (see Gilbert, this volume), some attributional processes (e.g., inferring that a favorable view of an
object is responsible for one’s seeking of it) likely require somewhat less effort than the active scrutiny of
attitude-relevant information. Because of this, self-perception might be used at times as a short cut for assessing
the validity of a stance toward an attitude object (perhaps so that effortful processing of information about the
object can be foregone). Various forms of the attributional approach have been studied in relation to the
attributions about attitude object itself, as well as attributions about the source of information about an attitude
object. Most generally, the attributional approach focuses on people inferring underlying characteristics about
themselves and others from the behaviors that they observe and the perceived situational constraints that are
imposed on those behaviors (e.g., Bem, 1965; Jones & Davis, 1965). Bem (1965) suggested that people
sometimes have no special knowledge of their own internal states and therefore must infer their attitudes in a
manner similar to that by which they infer the attitudes of others (i.e., from the observed behavior and context in
which it occurred).
During much of the 1970s, Bem’s self-perception theory was thought to provide an alternative account of
dissonance effects (Bem, 1972). For example, in the classic Festinger and Carlsmith (1959) study described
earlier, a person who observed an individual saying a task is interesting after being given $20 might infer that this
person liked the task less than a person who said it was interesting for only $1. If an external observer might
make this reasonable inference, so too might the person him or herself, Bem argued. Subsequent research
indicated, however, that self-perception processes could not account for all dissonance effects (e.g., Beauvois,
Bungert, & Mariette, 1995). Rather, the two processes operate in different domains. In particular, the underlying
internal discomfort mechanism of dissonance theory operates when a person engages in attitude-discrepant action
that is disagreeable (e.g., advocating a discrepant position in one's latitude of rejection; Fazio, Zanna, & Cooper,
1977; performing self-deprecating behavior; Jones, Rhodewalt, Berglas, & Skelton, 1981), whereas selfperception processes are more likely when a person engages in attitude-discrepant but more agreeable behavior
(e.g., advocating a discrepant position in one's latitude of acceptance; Fazio et al., 1977; performing a selfenhancing behavior; Jones et al., 1981). Also, self-perception processes do not require scrutiny of the central
merits of an attitudinal position (i.e., ?viewing” the behavior is all that is necessary), but dissonance reduction is
conceived as a thorough (but biased) form of information processing (Festinger, 1957). Thus, in elaboration
likelihood terms, one would expect self-perception processes to be more likely when elaboration likelihood is
low, but dissonance processes to be more likely when elaboration likelihood is high (e.g., the cognitions are very
important or self-relevant; Petty & Cacioppo, 1986a). Of course, especially difficult or complex attributions
might be less likely to occur under such low-effort settings.
Self-perception theory also accounted for some unique attitudinal phenomena. For example, the
overjustification effect occurs when a person is provided with more than sufficient reward for engaging in an
action that is already highly regarded (e.g., Lepper, Greene, & Nisbett, 1973). To the extent that the person
comes to attribute the action to the external reward rather than to the intrinsic enjoyment of the behavior,
attitudes toward the behavior will become less favorable (Deci, 1975). Thus, if people are provided with
extrinsic rewards for advocating a position that they already like, they may come to devalue the position when the
external rewards stop to the extent that they have come to view their attitude expression as caused by the rewards
rather than by the true merits of the position (e.g., Scott & Yalch, 1978).
As noted earlier, according to the ELM, people should be more likely to rely on simple self-perception
inferences when elaboration likelihood is low. Consistent with this notion, in one study, Taylor (1975) asked
women to evaluate pictures of men who they believed they would actually meet (high personal
relevance/elaboration), or not (low personal relevance/elaboration). Participants received false feedback about
their ?positive physiological responses” toward some of the men (Valins, 1966). The information about
physiological reactions influenced the women’s reported attitudes only when there was no expected meeting with
the man in the picture but not when the consequences were high (see also, Chaiken & Baldwin, 1981; Wood,
1982, for conceptually similar findings).
The attribution approach has also been useful in understanding how inferences about message sources
can influence attitudes. For example, Eagly, Chaiken, and Wood (1981) argued that people often approach a
persuasion situation with some expectation regarding the position a communicator will take. They argued that, if
the pre-message expectation is confirmed by the communicator's presentation, the recipient attributes the message
18
to the traits and pressures that generated the expectation rather than to the validity of the position espoused.
Thus, in these cases, the person needs to process the message to determine its validity. However, when the
premessage expectation is disconfirmed, the communicator is viewed as relatively trustworthy, the message as
veridical, and persuasion can occur without the need to process message arguments (e.g., Eagly, Wood, &
Chaiken, 1978). The reduced processing of trustworthy sources is especially evident among people low in need
for cognition (Priester & Petty, 1995). In addition, disconfirmation of a pre-message expectancy only leads to
perceiving the source as trustworthy when the disconfirmation entails the source violating his or her own selfinterest. If disconfirmed expectancies occur in the absence of a self-interest violation, then the surprise of the
disconfirmation enhances message processing (Harasty, Petty, & Priester, 1996).4
Heuristic/systematic model. Like the ELM, the heuristic/systematic model of persuasion (HSM)
considers multiple processes of persuasion. Importantly, the HSM identified a unique peripheral persuasion
process. Specifically, Chaiken (1980; 1987; Chaiken et al., 1989) proposed that in contrast to "systematic" (or
central route) processes, many source, message, and other cues are evaluated by means of simple cognitive
heuristics that people have learned on the basis of past experience and observation. Unlike attributional
inferences which can be novel and generated on-line, the use of heuristics was proposed to be dependent on their
availability and accessibility in memory.
According to the HSM, the likelihood of systematic processing increases whenever confidence in one's
attitude drops below the desired level of confidence (the "sufficiency threshold"). Whenever actual and desired
confidence are equal, heuristic processing is more likely. For example, because of prior personal experience,
people could base acceptance of a message on the expertise of the message source by retrieving the heuristic
"experts are usually correct” (Chaiken, 1980; Petty et al., 1981). Heuristics that are available in memory and
accessible (activated from memory) are the most likely to be used. Furthermore, the HSM holds that as the
motivation and ability to process increase, there is also an increased likelihood of heuristic processing. This is
because any factor that increases the importance of assessing the validity of support for a position should also
increase the salience (accessibility) of heuristics relevant to accomplishing that goal (Chaiken et al., 1989, p.
225). Although this enhancement of heuristic processing is hypothesized to occur, the increased scrutiny of the
merits of the attitude object might provide information that contradicts accessible heuristics, and thus might
attenuate the impact of heuristic cues (attenuation effect). As long as the two processing modes do not yield
conflicting reactions, however, increases in the importance of assessing attitude validity should increase the
impact of heuristic cues within this model (additivity effect).
The expanded HSM also deals with biased processing (Chaiken et al., 1989). According to this model,
bias can occur in at least two ways. First, Chaiken et al. argue that in some circumstances, heuristic processing
can bias systematic processing (e.g., accessing the heuristic that “experts are correct” can lead people to engage
in favorable elaboration of a message). Second, in addition to postulating an “accuracy motive” that produces
relatively objective information processing, two other motives (defense and impression management motives)
operate to produce biased processing either though biasing systematic or heuristic processing. When defense
motives are operating, for example, people are motivated to defend their existing attitudes but can do so either by
biased systematic processing or selectively using heuristics (Eagly et al., 1996).
Although considerable research supports the general predictions of the HSM (and typically the ELM as
well), little research has addressed the defining feature of the HSM -- the notion that people have learned and
stored in memory various persuasion heuristics that are used to evaluate a message. Some research has varied the
accessibility (Roskos-Ewoldsen & Fazio, 1992) or vividness (Pallak, 1983) of the peripheral cues in a message -and presumably the accessibility of the associated heuristics -- yet, relatively little evidence has been collected
relating to the accessibility of the heuristics themselves. In perhaps the most relevant research, Chaiken (1987)
had people memorize 8 phrases relevant to the length implies strength heuristic (e.g., "the more the merrier") or 8
irrelevant phrases. Memorizing the relevant phrases led to greater impact of a source claiming to have 10 versus
2 reasons than memorizing the irrelevant phrases. This effect occurred only for people classified as low in need
for cognition (Cacioppo & Petty, 1982) who would presumably be most likely to rely on a low effort heuristic
strategy. Cialdini (1987) analyzed several other heuristics that might be effective in influencing behavior (see
Cialdini, this volume). For example, although people might reason that the more people endorsing an object, the
better it is, they can also reason that the object is better if relatively few rather than many are available (i.e., the
19
"scarcity heuristic;" see Brock, 1968). Although many such cues have been shown to influence attitudes (mostly
when the elaboration likelihood is low), it has not been clearly demonstrated that such cues operate primarily
through stored heuristics rather than one of the other peripheral processes. Nevertheless, the heuristic notion has
itself served as a very useful heuristic for guiding interesting persuasion research.
Although the HSM and ELM share many features (e.g., attitude change can result from both high and
low effort mental processes which can be relatively objective or biased), a number of differences do exist
regarding the impact of cues in high-thought situations and mechanisms hypothesized to account for biased
processing outcomes. Recall that in the ELM, increased elaboration likelihood can lead to salient cues being
evaluated for strength just as the message arguments are evaluated (Petty, 1994). Thus, if a potential cue (e.g.,
attractive source) is scrutinized and found lacking (e.g., “it is biasing to go along just because he/she is
attractive”), then the cue can actually reduce persuasion if effortful processing leads to an overcorrection for the
perceived bias induced by the cue (see Wegener & Petty, 1997). If the cue is deemed relevant and informative
when scrutinized, however, then it would add to the impact of the other information. Note that in the ELM, this
additive impact is not a result of a low effort heuristic adding to the impact of high effort central/systematic
processing (cf., Maheswaran & Chaiken, 1991), but is due to the fact that the cue/heuristic is effortfully
scrutinized as a potential argument supporting the advocacy of the message.
Consider the notion of biased processing in the expanded HSM. In general, the idea that heuristic
processing can bias systematic processing is similar to the ELM multiple roles notion (Petty & Cacioppo, 1986a)
that variables that serve a cue (or heuristic) function when the elaboration likelihood is low can bias information
processing when the elaboration likelihood is high (though, in the ELM, the variable producing a bias need not
do so by invoking a heuristic; e.g. positive mood can make positive thoughts more likely to come to mind even if
a mood heuristic is not invoked). As noted previously, the ELM holds that objective processing occurs when
people have no a priori position to favor and that motivated biased processing (via either the central or the
peripheral route) can be produced when people prefer one position over another. Importantly, in the ELM,
people can come to prefer one position over another for a variety of motivational reasons (e.g., consistency,
reactance, self-esteem, etc.). Of most interest, accuracy motives can also lead to biased processing if people are
highly confident that their current view is correct and are motivated to defend it because of its presumed validity.
If people are confident that their attitudes are correct, any increase in the motivation to be accurate would
increase motivation to defend their attitudes. Conversely, reducing the need to be accurate would free people to
go along with others and abandon their presumably accurate views. Consider another way in which accuracy
motives can produce biased processing. If the elaboration likelihood is high and people are confronted with an
ambiguous argument, people may be more likely to elaborate it in a way that is favorable to the advocacy if the
source is perceived to be an expert than if the source is not (Chaiken & Maheswaran, 1994). This favorable
interpretation of the arguments could be motivated by wanting to have the most accurate interpretation of the
evidence and the assumption that an expert source is more likely to hold accurate opinions than a non-expert.
Alternatively, it might be that people prefer to identify with the position of experts for reasons of self-esteem
maintenance, or because of aversive feelings of imbalance or confusion if experts are wrong, and thus these
motives rather than accuracy could account for the biased processing. Thus, within the ELM framework,
accuracy motivation per se can bring about biased outcomes that look ?defensive,” and additional motivations
other than defense or impression management can be responsible for biased outcomes (in addition to the ability
factors discussed earlier).
Other Peripheral Processes
Mere exposure. When objects are presented to an individual on repeated occasions, this mere exposure
is capable of making the person's attitude toward the objects more positive (see Zajonc & Markus, 1982). An
early explanation of the mere exposure effect was provided by Titchener (1910), who proposed that familiar
objects led people to experience a "glow of warmth, a sense of ownership, a feeling of intimacy" (p. 411). Work
on this phenomenon has shown that simple repetition of objects can lead to more positive evaluations even when
people do not consciously recognize that the objects are familiar. For example, Kunst-Wilson and Zajonc (1980)
repeatedly presented polygon images under viewing conditions that resulted in chance reports of recognition.
During a later session, pairs of polygons were presented under ideal viewing conditions. In each pair, one shape
had been seen in the earlier session, but the other was new. When asked which shape they liked better and which
20
one they had seen before, research participants were unable to recognize beyond chance which polygon was new
and which was old, but they showed a significant preference for the "old" shapes.
Just because people do not consciously recognize that a stimulus was not presented many times
previously, does not mean that people are not aware at some level that these repeated stimuli are easier to
perceive and process. This perceptual fluency (Bornstein, 1989; Jacoby, Kelley, Brown, & Jasechko, 1989)
might be attributed to or confused with the favorability of the stimulus, but might also be attributed to other
stimulus dimensions (e.g., brightness, darkness, etc.), and account for why repeated exposure produces more
extreme judgments of a variety of stimulus-relevant dimensions (Mandler, Nakamura, & Shebo Van Zandt,
1987). In fact, if people attribute a sense of familiarity to the experimental procedure itself, the mere exposure
effect is weakened (Bornstein & D’Agostino, 1994). In a similar vein, Downing, Judd, and Brauer (1992) found
that larger numbers of repeated expressions of either evaluative or non-evaluative (i.e., color) responses led to
more extreme responses (on the evaluative or non-evaluative dimension, respectively) than fewer numbers of
expressions. Downing et al. (1992) reasoned that this occurred because of a link between the response and the
object becoming stronger over repeated expressions.
Mere exposure effects (especially on the favorability dimension) have been shown using a variety of
stimuli such as tones, nonsense syllables, Chinese ideographs, photographs of faces, and foreign words (e.g., see
Bornstein, 1989, for a review). It is important to note that all of these stimuli tend to be low in prior experience
and meaning to the people receiving them and thus are relatively unlikely to elicit spontaneous elaboration. In
fact, mere exposure appears to be especially successful in influencing attitudes when conscious processing of a
repeated stimulus is minimal (see Harrison, 1977) or impossible (e.g., when the stimulus is presented
subliminally, Bornstein & D'Agostino, 1992). Manipulations that increase thinking, such as evaluation
apprehension, reduce mere exposure effects (Kruglanski, Freund, & Bar-Tal, 1996; for additional discussion, see
Zajonc, this volume).
When more meaningful stimuli are presented, the effect of such exposures is quite different from the
usual mere exposure effect. That is, increased exposures of more meaningful stimuli enhance the dominant
cognitive response to the stimuli. Thus, attitudes toward negative words (e.g., "hate") and toward weak message
arguments actually become more unfavorable with increased exposures, but attitudes toward positive words (e.g.,
"love") and toward strong arguments become more favorable, at least until the point of tedium (e.g., Cacioppo &
Petty, 1989; Grush, 1976).
Balance. Although cognitive imbalance can lead to effortful (though potentially biased) scrutiny of
attitude-relevant information (Festinger, 1957), one can also conceive of less effortful means of addressing
imbalance. For example, consider the relations among a person, his or her boss, and an object (the three of which
make up a person-other-object triad). According to Heider (1958), balance occurs when people agree with people
they like (or with whom they are strongly associated) or when people disagree with people they dislike (or are
strongly dissociated). An imbalanced triad in which a person has a unit relationship with the other person (e.g.,
his or her boss), the person dislikes an object (e.g., a painting), but the boss likes the object, is likely to put the
person in an uncomfortable state (see Heider, 1958). Because the unit relationship is likely difficult to change,
the person might find that the easiest way to balance this triad is to change his or her view of the painting (see
Rosenberg & Abelson, 1960). This can be done simply because it feels “harmonious” and not because the person
has effortfully reconsidered the merits of the painting (as dissonance theory might suggest). In addition to
preferences for “balanced” relations among entities and attitudes, people have also been shown to prefer
“attraction” (i.e., positive sentiments between the two people in a triad) and “agreement” relations (i.e., when the
two people agree in their attitudes toward the object regardless of their sentiments toward each other; see Miller
& Norman, 1976). Importantly, such positivity preferences and any changes in attitudes that they might bring
about could take place with little or no consideration of what the person perceives to be the central merits of the
attitude object. One could also generate new relations among the elements in a triad by relating pairs of the
elements to new elements, but such processes might also take place without consideration of the central merits of
the attitude objects involved in the triad (see Insko, Songer, & McGarvey, 1974; Newcomb, 1968; for further
discussions).
Summary
Thus, a variety of attitude change processes have been discussed over the years, and these processes can
21
be generally classified according to the overall mental effort necessary for them to affect attitudes and according
to the role (or lack thereof) of scrutiny of the central merits of the attitude object (see Petty & Cacioppo, 1981).
Although many of these processes were initially introduced as global models of attitude change, contemporary
multi-process models emphasize the conditions under which the high- and low-effort processes are most likely to
occur. In the following sections, the multiple role feature of the ELM is used to organize and review empirical
work on persuasion variables. Incorporating traditional classifications of these variables, the presentation is
organized into source, message, recipient, and context categories.
PERSUASION VARIABLES
From Lasswell’s (1948) well-known question -- Who says what to whom with what effect? -- to Hovland
et al.’s, (1953) classic Communication and persuasion volume, to McGuire’s (1969, 1985)
communication/persuasion matrix model (described earlier), variables having an impact on attitude change have
traditionally been organized into source, message, recipient, and context categories. This organization scheme is
not very conceptual, is to some extent arbitrary (e.g., mood can be considered a “recipient” variable when
measured, but a “context” variable when manipulated), and breaks down as an organizational scheme when
categories of variables interact (e.g., source X message). Nonetheless, this organization has provided a handy
reference for those interested in selecting variables for applications of persuasion theory for over half a century.
Thus, in this section an overview and updating of some of the most researched variables falling into these
traditional categories is presented. First, however, it is useful to consider, in a general way at least, the variety of
processes by which any given variable might have its impact on persuasion.
The earliest research on source, message, recipient, and context variables tended to assume that the
particular variable under study (e.g., source credibility) had a unidirectional effect on attitude change (e.g.,
increasing credibility increased persuasion) and produced this effect by a particular process (e.g., increasing
credibility fostered learning of the message arguments). Decades of research on attitude change clearly indicated
that (unfortunately) these simple assumptions were untenable (Petty, 1996). The accumulated research indicated
that even simple and seemingly obvious variables like source credibility were sometimes associated with
increased influence (Kelman & Hovland, 1953) and sometimes with decreased influence (Sternthal, Dholakia, &
Leavitt, 1978). One solution to this complexity was to suggest that each outcome was caused by a different
psychological process. However, research on many variables indicated that the same process could sometimes
lead to opposite outcomes, and the same outcome could be caused by different processes (Petty & Cacioppo,
1986a).
Thus, in order to understand source, message, recipient, and contextual variables, it is necessary to relate
them to the underlying processes of persuasion. In the preceding sections of this chapter, processes that might be
responsible for changes in attitudes when the mental effort allocated was particularly high or low were discussed.
In addition, the factors of motivation and ability that determine whether the persuasion situation is characterized
by high or low amounts of mental effort were highlighted. Next, after reviewing the multiple processes by which
variables can have an impact on attitude change, the existing literature on the effects of source, message,
recipient, and context variables is examined, and the observed effects are related, when possible, to the multiple
roles possible for persuasion variables.
Multiple Roles for Persuasion Variables
As previously noted, across situations, people, and objects, there are differences in the extent to which
people are willing and able to put a high level of effort into arriving at their evaluations. To the extent that
people are both motivated and able to put effort into forming or changing their views of an object, they are likely
to carefully scrutinize information relevant to that object (i.e., they are likely to effortfully assess the “central
merits” of the object in order to determine the extent to which the object is good or bad; see Petty & Cacioppo,
1979b, 1986a). Thus, when motivation and ability are high, one way for a variable to influence judgments is for
it to be treated as an argument -- a piece of information relevant to determining the merit of the object or issue.
For example, if one wishes to assess the extent to which it would be good or bad to hire a particular person as a
model for a cosmetic product, then the appearance of that person is likely a central dimension to be used in that
assessment. Therefore, in such a case, variations in the attractiveness of the person could influence evaluations
through effortful consideration of the person’s attractiveness (along with considerations of all other information
relevant to determining merit such as past modeling experience).5
22
As discussed earlier, another way for a variable to influence judgments when motivation and ability are
high is to bias the processing of attitude-relevant information. That is, if multiple interpretations of information
are possible, a variable might make one interpretation more likely than other equally plausible interpretations.6
For example, it has been repeatedly shown that people assume that attractive people possess other positive traits
(e.g., Cooper, 1981; Thorndike, 1920). This halo effect could bias processing of information from an attractive
target by making positive interpretations of ambiguous information more likely than if the target were not
attractive. Thus, when motivation and ability to process attitude-relevant information are high, variables can
affect judgments by serving as arguments or by biasing interpretations of ambiguous attitude-relevant
information.
For some situations, people, or objects, however, either motivation or ability to process attitude-relevant
information is lacking. When this is the case, people are more likely to use some kind of short-cut for
determining what a reasonable view of the object might be (e.g., relying on one’s first impression rather than
scrutinizing all the information). That is, people low in motivation or ability are likely to form or change their
views of the object on the basis of some factor that allows them to do so without engaging in the cognitive work
needed to assess the central merits of the target object (e.g., Chaiken, 1980; Petty et al., 1981). Thus, when either
motivation or ability to process information is lacking, a variable can impact judgments if that variable can
influence attitudes by one of the relatively low effort processes that were described previously. People are
sometimes aware of the operation of these low effort processes, and sometimes they are not. In either case, they
arrive at a judgment of the object with relatively little mental effort expended.
Finally, some variables can affect attitudes by influencing one’s motivation or ability to think carefully
about judgment-relevant information. Of course, the likelihood of any given variable influencing the amount of
scrutiny is constrained by all of the background variables also affecting the baseline level of scrutiny. If the
baseline likelihood of elaboration is already quite low (e.g., because distraction is at a high level, Kiesler &
Mathog, 1968; Petty et al., 1976) or quite high (e.g., because the attitude object is very important or personally
relevant, Leippe & Elkin, 1987; Petty et al., 1981; or because the people receiving attitude-relevant information
are very high in need for cognition, Cacioppo et al., 1983), then impact of a variable on attitudes is most likely to
occur through the low- or high-elaboration roles outlined earlier. If background variables do not constrain
elaboration to be particularly high or low, and especially if a person is not sure whether or not effortful scrutiny
of information about the target is merited, however, then the variable might affect attitudes by helping to
determine the level of thought given to the merits of the attitude object.
In sum, according to the ELM, variables can influence judgments 1) by serving as arguments relevant to
determining the merits of an object or position, 2) by biasing processing of attitude-relevant information (both of
which are most likely when motivation and ability to scrutinize attitude-relevant information are high), 3) by
serving as a peripheral cue to the adequacy of a view of the target (when motivation or ability are low), and 4) by
itself affecting the level of scrutiny given to attitude-relevant information (when elaboration likelihood is not
constrained by other factors to be particularly high or low). Therefore, the ELM is a model of moderated
mediation (see Petty et al., 1993, for further discussion). It is important to note, however, that the ELM is not
aimed at predicting how every persuasion variable influences persuasion at each level of elaboration likelihood.
Rather, the ELM provides an organizing framework that specifies which classes of processes operate under
which levels of elaboration. Theoretical and empirical complements to the ELM framework specify the
particular effects of most variables (e.g., whether and when positive moods enhance versus reduce processing of
persuasive messages).
In the following sections, the ELM notion of multiple roles for variables across the elaboration
continuum is used to organize the work done on four broad classes of variables (i.e., source, message, recipient,
and context). Each section begins with a brief definition, and then the kinds of effects and general principles that
have been found within each class of variables are described.
Source Variables
Source variables refer to aspects of the person(s) presenting the persuasive appeal. At times, the identity
of a source might be made very explicit (e.g., when a candidate makes a speech on behalf of his or her
candidacy), but at other times, the source is merely implied (e.g., when an unpictured narrator announces a new
product, presumably on behalf of the company producing that product). Traditionally, source effects have been
23
organized according to a taxonomy introduced by Kelman (1958) which separated source factors into effects of
credibility, attractiveness, and power (see also McGuire, 1969). Kelman (1958) discussed credibility effects as
due to internalization, which entailed acceptance of information and integration of that information into one’s
existing cognitive system, attractiveness effects as due to identification, which relied on salience of one’s bond to
or relationship with the message source, and power effects as due to compliance, which would occur only as long
as the source maintained control over potential rewards and punishments. Although these classifications and
principles have proven quite useful, recent research suggests, for instance, that credibility can also influence
attitudes when relatively little information scrutiny or internalization takes place and that attractiveness has
effects even when no bond or relationship between source and message recipient is possible. An alternative
framework for organizing source effects focuses on the effects of sources across levels of elaboration likelihood,
with an eye toward the multiple roles that sources can play in persuasion settings. A variety of characteristics
related to source credibility, attractiveness, and power have been investigated.
Credibility
In one of the earliest investigations of source credibility, Hovland and Weiss (1951) presented students
with a message on one of four topics and then told them the source of the message (with the source either being
high or low in credibility). Although Hovland and Weiss (1951) focused on the overall effect of source
credibility (collapsed across messages), the credibility effect was more pronounced for the two topics that were
less likely to directly impact students (e.g., who to blame for a steel shortage) than for the two topics more likely
to be relevant to students (e.g., will TV decrease the number of movie theaters in operation). Thus, even early
studies of source credibility provided some indications that variables such as source credibility might not operate
in the same way (or to the same extent) in all circumstances. Although many studies of source credibility have
used sources that vary in both knowledge (expertise) and presumed honesty (trustworthiness; e.g., Kelman &
Hovland, 1953), a variety of studies have more directly studied the impact of differences in expertise or
trustworthiness per se (see Petty & Cacioppo, 1981).
Expertise. Perhaps the most prevalent characterization of source expertise effects is as a peripheral cue.
Consistent with the peripheral cue notion, highly expert sources have led to more persuasion than inexpert
sources to a greater extent when a topic is presented as low rather than high in personal relevance (e.g., Petty et
al., 1981) or when a topic has little rather than great direct relation to recipients’ lives (e.g., Rhine & Severance,
1970). Similarly, source expertise effects have been moderated by other variables thought to influence the
amount of scrutiny given to persuasive messages. For example, source expertise has a greater impact when
distraction is high rather than low (e.g., Kiesler & Mathog, 1968), when topic relevant knowledge is low rather
than high (Wood & Kallgren, 1988), and when messages are externally paced (i.e., taped) rather than self-paced
(i.e., written; e.g., Andreoli & Worchel, 1978).
Although less commonly studied, source expertise has also been shown to relate to the amount of
scrutiny given to persuasive messages when the elaboration likelihood is not constrained at a high or low level.
Specifically, Heesacker, Petty, and Cacioppo (1983) found that field-dependent message recipients (see Witkin,
Goodenough, & Oltman, 1979) engaged in greater message scrutiny (i.e., were more persuaded by strong than by
weak arguments) when the message was presented by an expert rather than inexpert source. DeBono and Harnish
(1988) found that individuals low in self-monitoring (Snyder, 1974) processed messages to a greater extent when
they were presented by an expert rather than an attractive source (see additional discussion in recipient variable
section). Moore, Hausknecht, and Thamodaran (1986) established three levels of elaboration likelihood by
presenting an advertisement at a fast, moderately fast, or normal speech rate and manipulated source expertise
and argument quality. When the ad was presented rapidly, recipients were influenced by source expertise, but
not the quality of the arguments (peripheral cue effect). When the ad was presented at a normal pace, the impact
of argument quality was increased and the source expertise effect decreased compared to the rapid pace. These
effects replicate other research on the tradeoff between cue and argument effects across the elaboration
continuum. Of greatest interest, when the ad was presented at a moderately fast pace such that processing was
possible but challenging, expertise interacted with argument quality such that the message received greater
scrutiny when presented by an expert than a non-expert source.
Finally, the potential biasing impact of source expertise under high elaboration conditions was
demonstrated by Chaiken and Maheswaran (1994). First, they showed that source expertise had a greater impact
24
on attitudes when unambiguous strong or weak arguments were presented on an unimportant topic rather than an
important topic (similar to Petty et al., 1981). More importantly, Chaiken and Maheswaran (1994) also showed
that an expert source was more persuasive than an inexpert source under both high and low importance
conditions when the arguments were ambiguous (i.e., not clearly strong or weak). When the ambiguous
arguments were presented on an important topic, expertise significantly affected the valence of message-relevant
thinking (i.e., biased message processing), but when the topic was unimportant, expertise did not affect messagerelevant thoughts (i.e., expertise acted as a persuasion cue).
Trustworthiness. Credible sources are not only knowledgeable (i.e., expert) about the topic, but are also
perceived as trustworthy (i.e., likely to tell the truth; see Hass, 1981; Petty & Cacioppo, 1981). In fact,
trustworthiness per se has been shown to have effects on overall persuasion and on processing of persuasive
messages. For example, as noted earlier, Eagly and her colleagues found that sources regarded as trustworthy or
sincere (because of presenting a point of view that disagreed with the views of a message audience) were more
persuasive than sources perceived as untrustworthy (e.g., Eagly, Wood, & Chaiken, 1978; Mills & Jellison,
1967). Trustworthiness is a clear cue to validity and if the message does not warrant processing for other
reasons, it can simply be accepted if the source is presumed to be knowledgeable and trustworthy. Priester and
Petty (1995) found that people who do not enjoy thinking were especially likely to accept a message from a
trustworthy source without scrutiny. When the source was untrustworthy, however, people low in need for
cognition (Cacioppo & Petty, 1982) engaged in as much message processing as those high in need for cognition
(because the untrustworthy character of the source left them uncertain whether or not the position of the message
held merit or not).
Attractiveness/Likeableness
Although some manipulations of ?attractiveness” have varied physical characteristics of the source (e.g.,
Snyder & Rothbart, 1971), many manipulations of this construct have also incorporated likeable versus
dislikeable features of the source (e.g., Petty et al., 1983; Zimbardo et al., 1965). In addition, when physical
characteristics have been manipulated, the effects of physical attractiveness have appeared to be mediated by
effects on liking of the communicator (see Chaiken, 1986).
As with other source characteristics, attractiveness or likeableness has been studied primarily in relation
to its role as a peripheral cue (see Chaiken, 1987; Cialdini, 1987). Consistent with the cue notion, effects of
source attractiveness/likeableness have been greater when elaboration likelihood is low rather than high. That is,
source attractiveness or liking has been observed to exert a greater impact when relevance of the topic is low
rather than high (e.g., Chaiken, 1980; Petty et al., 1983), when attitude relevant knowledge is low rather than high
(e.g., Wood & Kallgren, 1988), and when messages are externally paced on audio or video tapes rather than selfpaced and written (e.g., Chaiken & Eagly, 1983). In each of these experiments, arguments that were
unambiguously strong or weak have been used, but if ambiguous arguments had been used, one might also have
found evidence of biased processing when elaboration was high.
In addition, the attractiveness of the source has been shown to influence the amount of message scrutiny
that takes place when elaboration likelihood is moderate or ambiguous. For example, Puckett, Petty, Cacioppo,
and Fisher (1983) presented college students with a message advocating comprehensive exams as a graduation
requirement, but the time frame for any consideration of the issue at their university was left unspecified
(rendering the personal relevance ambiguous). Puckett et al. (1983) found that message recipients engaged in
greater message scrutiny when the message was presented by a socially attractive rather than unattractive source.
Also, DeBono and Harnish (1988) found that individuals high in self-monitoring (Snyder, 1974) processed
messages to a greater extent when they were presented by an attractive rather than an expert source (see also
discussion in section on recipient variables).
In a study examining multiple roles for source attractiveness, Shavitt, Swan, Lowery, and Wänke (1994)
manipulated the attractiveness of an endorser in an advertisement for a restaurant, the salient (central) features of
the product (either unrelated to attractiveness -- taste and aroma -- or related to attractiveness -- public image of
the restaurant), and motivation to process the ad. When endorser attractiveness was unrelated to the central
merits of the product (and the ELM would predict that any impact of attractiveness would be due to its impact as
a peripheral cue), attractiveness had an impact on evaluations of the product under low but not high motivation
(and had little impact on thoughts about the product). However, when endorser attractiveness was related to the
25
central merits of the product (and thus could itself act as an argument), the same variation in attractiveness
influenced evaluations of the product under high but not low motivation (and under high motivation, influenced
the favorability of cognitive responses to the ad).
Power
Power of the source over a message recipient has been analyzed in terms of the extent to which the
source is perceived as having control over positive or negative sanctions, concern over whether the recipient
complies with the request, and the ability to monitor whether or not the recipient accepts the source’s position or
not (see McGuire, 1969). Each of these aspects of power have been shown to affect the persuasiveness of
sources (with powerful sources persuading others more than weak sources; e.g., Festinger & Thibaut, 1951;
Raven & French, 1958).7 Yet, little work has investigated the information processing consequences of message
presentation by a powerful versus weak source. In some instances, change to powerful sources undoubtedly
reflects mere compliance rather than a true expression of opinion (Kelman, 1958). However, it is possible that as
long as the power of the source is not obviously coercive, powerful sources could induce genuine change because
of the operation of low-effort or high-effort processes.
One area of research with direct implications for these processes is the work on stereotyping of and by
powerful people (e.g., Fiske, Morling, & Stevens, 1996). For example, when people depend on the actions of
others to reap benefits, those people form more detailed, individuated impressions (as opposed to simple,
category-based impressions -- stereotypes) of the others (Fiske & Neuberg, 1990). This occurs not only in
situations where people are mutually interdependent, but also when asymmetrical power exists over tangible
rewards (i.e., one person depends on the other, but not vice versa). Interestingly, Fiske et al. (1996) argued that
when one person possesses evaluative power over another rather than power over task outcomes, processing of
information about the powerful individual is positively biased rather than accurate, presumably because people
want to convince themselves that the evaluator will be generous. This work suggests that persuasive messages
presented by a source with power over task outcomes might receive greater scrutiny than the same messages
presented by a powerless source (so long as the elaboration likelihood is not already constrained by other
factors), but that messages from sources with power over evaluations might be more likely to be processed in a
biased fashion.
Each of the multiple roles outlined by the ELM would seem applicable to sources that vary in power. If
processing is limited (e.g., by distraction), one might agree with a powerful source simply because it is generally
good to agree with people who control one’s fate. If the power of a source makes it appear more likely that a
proposed policy will be enacted, it could be that people would be more likely to process information about that
policy than if the source lacked the ability to institute the program. Also, if thinking is already extensive (e.g.,
because the topic is of great intrinsic interest or relevance), one might be disposed toward adopting
interpretations of the information that agree with a powerful source. Or, if the powerful source induces reactance
(e.g., from attempting overt coercion), motivation to counterargue (privately, at least) might be provoked.
Additional Source Factors Related to Credibility, Liking, and Power
A number of source characteristics have been studied that appear at least somewhat related to perceptions
of credibility, liking, or power. That is, sources belonging to different groups can speak at a certain speeds, in a
certain styles, using certain forms of language, and message recipients often make inferences about the
credibility, likeableness, or power qualities of the source based on these variations. Thus, each of these “style”
variables can potentially have an effect on message processing and persuasion at least in part because of their
effects on perceptions of the source’s credibility, likeableness, power, or some combination of the three.
Speed of speech. Several studies of the rate (i.e., words per minute) with which speakers present their
position have shown a relation between speech rate and judged credibility of the source (i.e., speaking quickly
has been associated with greater perceived credibility; Miller, Maruyama, Beaber, & Valone, 1976). This effect
has not been universal, however (see O’Keefe, 1990). One reason for the lack of consistency in relations
between speech rate and judged credibility might be that speech rate has also been found to influence elaboration
of persuasive messages. As noted earlier, Moore et al. (1986) found that rapidly presented radio ads led to a
reduced effect of argument quality than when the same ads were presented at a more moderate speed. Smith and
Shaffer (1995) found that persuasive effects of faster speech were mediated by perceptions of credibility when
the message was low to moderate in personal relevance, but faster speech had no direct effect on credibility or
26
persuasion for a message high in personal relevance (though faster speech rate did decrease processing of the
messages when personal relevance was high -- presumably because it was more difficult to process the message
when it was presented quickly).8
Demographic variables. One can imagine a host of demographic variables being related to perceptions of
likeableness, credibility, or power. Although many demographic source characteristics have been studied (e.g.,
gender, age, and ethnicity), few of these studies have considered the possible simple versus elaborative nature of
such effects. To the extent that effects of these variables occur because of inferences related to credibility,
likeableness, or power, demographic variables might serve the multiple roles outlined earlier for these variables.
For example, demonstrating a classic effect of source gender, Goldberg (1968) found that the scientific
content of an article labeled as written by ?John McKay” was rated more favorably than when the same article
was attributed to ?Joan McKay.” Why did this occur? Depending on one’s assumptions about the level of
scrutiny given to the material in the article, there could be a variety of reasons (see Deaux & LaFrance, this
volume). If scrutiny was relatively low, the male or female label could have acted as a simple persuasion cue.
The conceptual reason for this cue impact could be, for example, that men were viewed as more expert for the
particular topic involved (i.e., science). Such a possibility would also account for situations in which an article
written on a traditionally feminine topic is rated more favorably when attributed to a female rather than male
author (e.g., Levenson, Burford, & Davis, 1975). If scrutiny of the information is high, the effects could be
because of bias in favor of the author whose gender matches the topic or against the author whose gender
mismatches the topic. It could also be, however, that gender of the source influenced how much the information
in the article was processed (especially if background levels of elaboration were relatively moderate, and people
were unsure whether the article merited much scrutiny). Although little empirical work has attempted to
explicate the processes responsible for source gender effects, such effects might include each of the multiple
roles outlined earlier. Paying greater attention to the factors determining elaboration likelihood in such studies
might also help to account for the heterogeneity of effects across studies (see a recent meta-analysis by Swim,
Borgida, Maruyama, & Myers, 1989).
People of different ages are also likely to be viewed as differing in expertise (or in trustworthiness,
power, or likeableness) on many topics, as would people of different ethnicities or socio-economic status (e.g.,
consider the perceived self-interest of well-to-do members of congress arguing on behalf of tax breaks for the
wealthiest segments of society). To the extent that these inferences take place, demographic variables such as
age, ethnicity, or socio-economic status of the source should also assume the multiple roles taken on by the
conceptual variables of credibility, likeableness, or power. Some work associating demographic characteristics
with one or more of the multiple roles has taken place. For example, early work suggested a cue effect for source
race (e.g., Whittler, 1984), but more recent work has studied race as it relates to processing of a persuasive
message. White and Harkins (1994) found that Caucasian message recipients process information presented by
an African American source more than when the same information was presented by a Caucasian source.
According to White and Harkins, this occurred because message recipients held a negative attitude toward the
African American social group, but did not want to appear racist (see also Fazio, Jackson, Dunton, & Williams,
1995; Gaertner & Dovidio, 1986). Consistent with this notion, White and Harkins (1994) also found more
extensive processing of messages presented by a member of another group toward which message recipients held
a negative attitude (i.e., Hispanics) than for messages presented by sources from other equally novel groups
toward which message recipients were not as negative (i.e., Asians and Native Americans). This aversive racism
view of ethnicity and message processing might also suggest that one could find biases in processing of
ambiguous messages presented by African American or Hispanic sources. The multiple roles of these and other
demographic variables should receive attention in the future.
Majority/minority status. Like early work on many persuasion variables, initial studies of
majority/minority source status tended to ask questions such as whether majorities or minorities had greater
influence (e.g., Asch, 1956; Moscovici, Lage, & Naffrechoux, 1969). Over the years, the primary research
question has changed to whether majorities and minorities bring influence through the same or different
persuasion processes (Moscovici, 1980; Nemeth, 1986). It has also been acknowledged recently that the
majority/minority status of a source can have various effects depending on the motivational and cognitive factors
present (see Kruglanski & Mackie, 1990; Mugny & Perez, 1991; and Wood, Lundgren, Ouellete, Busceme, &
27
Blackstone, 1994; for reviews of minority influence work).
Within the multiple roles perspective, when the elaboration likelihood is low, majority/minority source
status is most likely to serve as a simple cue. Thus, whether a position is endorsed by a majority or a minority of
others can provide message recipients with a simple decision rule as to whether they should agree with the
message especially when people have no special interest or knowledge about the issue, or no message is even
presented (Giner-Sorolla & Chaiken, 1996; Petty & Cacioppo, 1986a). These effects are presumably due to
inferences about the presumed validity of the position or assumptions about the credibility, likeableness, or
power of the source associated with the majority or minority view. Source perceptions do not necessarily covary
directly with majority/minority status (Kruglanski & Mackie, 1990), however. For example, consistency in a
source's behavior can increase attributions of competence (e.g., Moscovici & Neve, 1973) or can increase
attributions of rigidity (e.g., Levine, Saxe, & Harris, 1976; see Maass & Clark, 1984).
When people are unsure whether they should carefully scrutinize the message or not, majority-minority
status can determine the amount of message scrutiny. Interestingly, some researchers have concluded that
majority sources foster greater elaboration, but others have concluded that minority sources foster greater
elaboration. For example, Mackie (1987) found that message recipients tended to recall more majority than
minority arguments, and also generated more favorable cognitive responses to the majority message. Based on
these findings, she concluded that majority sources induce greater message processing that is biased in a
favorable direction. However, other researchers concluded that minorities foster higher levels of elaboration
(e.g., Maass & Clark, 1983). More recent research has found that either majorities or minorities can induce
greater message scrutiny depending upon other factors in the persuasion situation. For example, Baker and Petty
(1994) found that people engaged in greater scrutiny of a counterattitudinal message when it was portrayed as the
majority position, but more scrutiny of a proattitudinal message when it was portrayed as the minority position.
They reasoned that if people receive a counterattitudinal message from a majority source, this implies that the
message recipient is in the minority which can be surprising and perhaps even threatening, making scrutiny of
what the majority has to say more likely. If a recipient encounters a proattitudinal message, however, the same
reasoning predicts greater processing when the source is a minority because a proattitudinal message from a
minority implies that the message recipient is also in the minority (which can be surprising or threatening). In a
somewhat similar study, Giner-Sorolla and Chaiken (1996) presented students with poll results that either
supported or opposed their recently stated vested interests (rather than their attitudes per se). In this study,
enhanced message processing was evident when the poll was incongruent with students’ vested interests over
when the poll was congenial. Although the students did not report being surprised by the poll, they may well
have been threatened by it.
Finally, when the motivation and ability to process an incoming message are high, majority/minority
status should impact persuasion primarily by influencing the nature of thoughts that come to mind. For example,
Mackie (1987) noted that in her research, biased processing might have produced the relatively greater proportion
of positive cognitive responses generated for messages from the majority source. Trost, Maass, and Kenrick
(1992) found that recipients for whom the message was highly relevant derogated the minority message more
than the majority communication. Thus, biased information processing of majority and minority sources might
be found primarily under conditions that foster elaboration.
Similarity to receiver. A great deal of work suggests that people like other people with whom they share
similar attitudes (e.g., Byrne & Griffitt, 1966) or ideology (Newcomb, 1956) and dislike those with whom they
disagree (e.g., Rosenbaum, 1986). This source-receiver similarity has also been shown to increase persuasion
(e.g., Brock, 1965). Sometimes there are benefits of dissimilarity, however. For example, although agreement
from similar sources induces greater confidence in one’s judgment when the issue is perceived to be a subjective
one, agreement from dissimilar sources induces greater confidence when the issue is perceived as objective
(Goethals & Nelson, 1973; see also Crano & Hannula-Bral, 1994). Unfortunately, little work has investigated the
mechanisms by which persuasive effects of similarity might occur. As with the other source variables, similarity
could produce attitude change by acting as a simple persuasion cue when elaboration is low, by biasing
processing when elaboration is high, or by affecting the amount of processing when background levels of
elaboration likelihood are relatively moderate.
One exception to this dearth of process-level investigation relates to the effects of messages presented by
28
an in-group versus out-group source (i.e., by a person who is similar to the message recipient by virtue of
belonging to the same group as the recipient or is dissimilar by belonging to a different group). Mackie, Worth,
and Asuncion (1990) presented University of California-Santa Barbara research participants with strong or weak
messages on either a topic presumably irrelevant to them (i.e., acid rain in the northeast) or presumably relevant
(i.e., oil drilling off the southwest coast of the U.S.). The messages were attributed to either a UCSB student
(i.e., an in-group member) or to a student of the University of New Hampshire (i.e., an out-group member).
When people received the message on the irrelevant topic, the in-group source was more persuasive than the outgroup source, regardless of the quality of the arguments presented. That is, when the message was low in
relevance, group membership of the source acted as a simple persuasion cue. When the message was on the more
relevant topic, however, message recipients processed the messages presented by the in-group source more than
when the same messages were presented by the out-group source. Thus, for the more group-relevant message,
similarity based on group membership influenced the amount of processing of the persuasive messages.9
Number of Sources
In addition to particular characteristics of sources, some research attention has been given to the sheer
number of sources who present the arguments in a persuasive message. Although early work focused on
conformity pressures (e.g., Asch, 1951) or on agreement based on a desire to appear correct (see Jellison &
Arkin, 1977), additional evidence supports the view that multiple sources can influence scrutiny of message
arguments. Harkins and Petty (1981a) hypothesized that when each argument in a message is presented by a
different source, elaboration of the message content is enhanced (see also Moore & Reardon, 1987). In addition,
this effect occurs if reception of the message is the person’s only task, but is eliminated if recipients are
distracted by a secondary task (e.g., Harkins & Petty, 1981b).
Why do people process information from multiple sources more than when the same information is
presented by one source? It appears that this effect is due to perceptions that the sources represent independent
perspectives (and thus, the converging view is more worthy of consideration). Evidence consistent with this
notion includes undermining the usual multiple source processing effect by presenting the sources as part of a
committee that worked together to generate the arguments (Harkins & Petty, 1987). Importantly, this method of
undermining the processing of a message presented by multiple sources is only effective if it is introduced before
rather than after message presentation. Also, if multiple sources are described as being similar rather than
dissimilar in background, the multiple source effect is reduced (Harkins & Petty, 1987).
These studies have been conducted without specifying the relevance of the communication for message
recipients. If personal relevance of the topic were extremely low (or elaboration likelihood were constrained to
be low by some factor such as distraction), it seems likely that the mere number of sources would influence
persuasion by a low-effort peripheral process. It is also conceivable that the number of sources could bias the
thoughts that people have when elaboration likelihood is high (especially for message recipients who care about
agreeing with significant others, i.e., high self-monitors).
Message Variables
Message variables refer to aspects of the communication itself. At a minimal level, a persuasion
situation contains some topic (e.g., capital punishment) or attitude object (e.g., ice-cream) that is the focus of the
influence attempt. Usually, the message has many more features, but sometimes the attitude object alone serves
as the “message” when it is simply repeated (e.g., Zajonc, 1968) as described in our previous discussion of mere
exposure effects. In addition to presenting the topic of the message, the message usually takes a particular
position, includes some substantive reasons supporting the position taken, and can be organized in a variety of
ways.
Message Topic, Position and Style
Issue-relevance/importance. Perhaps the most investigated aspect of the message topic is its importance
or relevance to the message recipient. This feature of a message has been referred to as issue-involvement (e.g.,
Zimbardo, 1960), personal involvement (Thomsen et al., 1995), personal and self-relevance (Petty, Cacioppo, &
Haugtvedt, 1992), vested interest (Crano, 1995), attitude importance (e.g., Boninger, Krosnick, Berent, &
Fabrigar, 1995), and ego-involvement (Sherif, Sherif, & Nebergall, 1965). Although the terms are different, the
basic notion is that there are some issues that people care about more than others. The primary determinant of
how much a person cares about some issue is the extent to which the issue is relevant to some aspect of oneself
29
(i.e., one’s beliefs, possessions, values, groups, etc., Boninger et al., 1995; Petty et al., 1992). In this regard,
issue importance can be considered a “recipient” variable, but because the influence agent can do things to the
message to enhance its perceived personal relevance (e.g., using personal rather than impersonal pronouns;
Burnkrant & Unnava, 1989), it is discussed here.
Early analyses of the relevance or importance of the topic were based on the social judgment theory
notion that one’s attitude served as a stronger judgmental anchor when the topic was ego-involving (Sherif et al.,
1965). This meant that an involving proattitudinal message would be seen as closer to one’s own position than
an uninvolving one (i.e., involvement led to greater assimilation) and an involving counterattitudinal message
would be seen as further away from one’s own position than an uninvolving one (i.e., involvement led to greater
contrast). According to social judgment theory, both judgmental distortions should reduce the likelihood of
attitude change. This is because discrepancy was thought to foster attitude change for agreeable communications
(and perceptions of discrepancy are decreased with assimilation) but discrepancy inhibited change for
undesirable communications (and discrepancy was perceived as greater with greater contrast; see subsequent
discussion of message discrepancy).
In contrast to this view, Petty and Cacioppo (1979b) postulated that increased personal relevance or
importance would enhance thinking about the communication which would increase persuasion if the arguments
were strong, but decrease persuasion if the arguments were weak. In a meta-analytic review of research on issueinvolvement, Johnson and Eagly (1989) concluded that this prediction was supported, but with qualifications.
Specifically, they argued that when studies examined whether the issue concerned an important consequence for
an individual or not (“outcome-relevant involvement”), involvement interacted with argument quality as
predicted. However, when studies examined whether the issue concerned participants’ cherished beliefs or not
(“value-relevant involvement”), two effects were observed. In addition to the interaction of involvement with
argument quality, a main effect of involvement was observed such that increasing involvement led to less
persuasion as predicted by social judgment theory. Petty and Cacioppo (1990) argued that because the latter
studies tended to be correlational rather than experimental, involvement was likely to be confounded with a
number of other variables. That is, participants classified as high in value relevance were likely to have more
extreme attitudes, more knowledge about their attitudes, and so forth, than did participants who were classified as
low in value relevance. Thus, the self-relevance per se would increase message processing (accounting for the
interaction of involvement and argument quality), but the more polarized attitudes and greater issue-relevant
knowledge on high value-relevant issues would bias the processing in an attitude-congruent direction (accounting
for the main effect of involvement).
A similar argument can be made regarding recent work on attitude importance. As in the early work on
social judgment theory, work on importance tends to find that people are more likely to resist a messages when
they consider the topic to be high rather than low in importance (e.g., Zurwerink & Devine, 1996; see Boninger et
al., 1995, for a review). Again, it seems likely that this resistance is not necessarily due to issue importance per
se. That is, there is no reason to reject a message simply because the topic is personally important. Rather, just
as in the social judgment research on ego-involvement, investigators interested in importance have measured this
construct rather than manipulated it. Thus, importance is likely to be confounded with a number of other
variables such as attitude extremity, knowledge, commitment, etc., and it is these constructs that could be
responsible for the resistance. Consistent with this notion, when an unfamiliar issue is made more important by
introducing it as likely to personally affect the message recipient, increases in objective processing of the
message arguments are the most likely result (see Petty et al., 1992). That is, importance influences the extent of
information processing, but it is the other variables (many of which are likely related to attitude strength) that
produce both cognitive and affective biases in the ongoing processing activity (see also Wegener, Downing,
Krosnick, & Petty, 1995, for discussions about manipulations and measures of this and other strength constructs).
In addition to these information processing consequences, there are other important implications of
making salient a link between a position and the message recipient. For example, people tend to like things that
are associated with themselves more than things that are associated with others. Thus, research on dissonance
theory showed that objects tend to be seen as more valuable as soon as an individual chooses it (e.g., Brehm,
1956). In fact, objects are also seen as more valuable even if people are simply given the items and no choice is
involved (Kahneman, Knetsch, & Thaler, 1991). People overvalue members of their in-group (Tajfel, 1970), and
30
have even shown preferences for the letters in their own names over other letters (Nuttin, 1985). Recall that one
explanation for why active generation of arguments produces more persuasion than passive exposure to them is
that people find their own arguments to be superior (Greenwald & Albert, 1968). This preference for things
associated with the self has been called the “ownness bias” (Perloff & Brock, 1980), the “mere ownership” effect
(Nuttin, 1985), and the “instant endowment” effect (Kahneman et al., 1990). Thus, if a speaker presented a
message about an object that a person just purchased, the person would presumably be especially motivated to
see the merits of the object. If the object was sufficiently important, the self-linkage would presumably induce
biased processing. If the object was unimportant or processing was impaired, people might simply reason that “if
it’s mine, it must be good.” Research in which self-linkage produced greater processing rather than biased
processing has tended to make salient the potential relevance of the object or issue to the self (e.g., you have the
option of choosing this product; Petty et al., 1983) rather than the certain relevance (e.g., you will receive this
product; see Petty & Cacioppo, 1990).
Position/Discrepancy. Perhaps the most salient initial feature of a message is whether it takes a position
that the recipient generally finds to be agreeable (proattitudinal message) or disagreeable (counterattitudinal
message). A number of investigators have proposed that attitude change should be an increasing function of
message discrepancy (how far the message position is from one’s own attitude; e.g., Anderson & Hovland, 1957;
Hunter, Danes, & Cohen, 1984). Social judgment theorists (e.g., Sherif & Hovland, 1961) provided a more
complex hypothesis. The prediction was that attitude change was an increasing function of discrepancy as long
as the message took a position in the recipient’s latitude of acceptance (i.e., the range of positions the person
found agreeable), but was a decreasing function of discrepancy when the message took a position in the latitude
of rejection (the range of positions the person found objectionable). Attitude change was proposed to reach a
peak when the message took a position in the latitude of non-commitment (the range of positions between the
latitudes of acceptance and rejection). Although many studies found an overall inverted-U pattern between
message discrepancy and attitude change as expected by the theory (e.g., Hovland, Harvey, & Sherif, 1957) -especially when the source was of low credibility (e.g. Aronson, Turner, & Carlsmith, 1963) -- careful analyses
did not prove congenial to the view that recipients’ latitudes moderated the effects as expected (e.g., Eagly &
Teelak, 1972).
Alternatives to the social judgment theory analysis of message position have been sparse, though a few
suggestions have been made. According to the ELM (Petty et al., 1992), for example, message discrepancy could
serve in several roles. Specifically, when the elaboration likelihood is low, message position could serve as a
simple cue. That is, agreeable sounding messages would be accepted but disagreeable sounding messages would
be rejected with relatively little scrutiny. When the elaboration likelihood is moderate, the message position will
determine, in part, the extent of message processing. For example, counterattitudinal messages (which threaten a
person’s views) might often receive greater scrutiny than proattitudinal messages (Cacioppo & Petty, 1979b;
Edwards & Smith, 1996). When the elaboration likelihood is high, people will engage in negatively biased
processing of counterattitudinal messages, but in positively biased processing of proattitudinal communications.
Zanna (1993) further suggested that the bias will be greater for counter than for pro communications. That is,
biased processing need not be symmetric -- people can be more biased in their assessment of messages that
disagree than those that agree with them.
Although the multiple roles for message position have not been examined in research, some evidence is
consistent with the proposition that counterattitudinal messages receive greater scrutiny than proattitudinal
messages when other factors have not constrained the elaboration likelihood to be high or low. In one study, for
example, recipients generated a greater number of counterargument thoughts as the message became more
counterattitudinal (Brock, 1967). In another study, students were presented with one set of arguments that were
framed as supporting either a pro or a counterattitudinal position (Cacioppo & Petty, 1979b). Recipients recalled
more of the arguments when they were used in support of the counterattitudinal advocacy suggesting that this
framing induced greater attention to the message content (see also Worth & Mackie, 1987). In other research,
students who were given an unfavorable medical diagnosis engaged in greater thought about it than did students
who were given a favorable medical diagnosis (i.e., they generated more alternative explanations for the
counterattitudinal than the proattitudinal information; Ditto & Lopez, 1992). As might be expected if people
scrutinize counterattitudinal information more than proattitudinal information, people take longer when
31
processing the latter than the former Edwards & Smith, 1996). Of course, as the issue becomes more important
and emotionally involving, the enhanced processing of counterattitudinal messages should also be biased in an
unfavorable direction.
Conclusion drawing. Another issue that has generated a small body of research is the question of
whether the message should make its position explicit or whether the specific point being advocated should be
implicit. For example, a college president can provide arguments about the benefits of a tuition increase without
ever explicitly stating the conclusion that tuition should be raised by 10%. A number of studies have suggested
that it is preferable to make the message position explicit (e.g., Hovland et al., 1949), though other research
suggests that if the recipient draws the conclusion on his or her own, this can be superior (Fine, 1957). The
problem seems to be that recipients are often either unable to draw the correct conclusion or are unmotivated to
do so (McGuire, 1969). However, when people are motivated and able to draw the conclusion for themselves,
such as when the message is highly involving or the recipients are high in their propensity to think, then it can be
better to leave the conclusion implicit (e.g., Stayman & Kardes, 1992). Engaging in the work of self-generation
is also likely to make the conclusion more memorable (Slamecka & Graf, 1978).
Use of rhetorical questions. Although the most common form of presenting an argument is to make
statements, message arguments can include rhetorical questions that ask whether the argument is true. For
example, one could summarize an argument with a statement that “institution of comprehensive exams will aid
students who are applying to graduate schools”, or one could ask “wouldn’t institution of comprehensive exams
aid students applying to graduate schools?” (e.g., Petty, Cacioppo, & Heesacker, 1981). Some research suggests
that inclusion of rhetorical questions makes the speaker appear more polite and likeable (e.g., Bates, 1976). As
such, use of rhetorical questions could function as a source cue when the elaboration likelihood is low, or bias
processing when the elaboration likelihood is high. At times, however, rhetorical questions also make the
speaker appear less confident (Newcombe & Arnkoff, 1979) which could operate as a negative rather than
positive cue (or bias). In addition, Zillmann (1972) suggested that use of rhetorical questions becomes associated
through socialization with strength of arguments (i.e., people only tend to use rhetoricals when arguments are
strong). Therefore, the presence of rhetorical questions could itself be used as a signal that the information is of
high quality.
As noted earlier, Petty et al. (1981) proposed that the use of rhetorical questions in a message could
increase message processing if people were not ordinarily inclined to think about the communication. Use of a
rhetorical question literally requires the person to think about the argument just presented in order to address the
question. A number of recent studies have supported this proposal (e.g., Burnkrant & Howard, 1984; Petty et al.,
1981). Specifically, when messages are audio taped, it has been found that use of rhetorical questions enhances
processing when baseline levels of elaboration are low, but use of rhetorical questions disrupts processing when
baseline levels of elaboration are high (Petty et al., 1981). One possible reason for the latter effect is that the
rhetoricals are distracting when people are ordinarily inclined to think about the message. Consistent with this
distraction notion, the presence of rhetorical questions only disrupts processing when the message is externally
paced. When the message is self-paced and when rhetorical questions precede rather than follow the arguments,
rhetorical questions increase message processing (e.g., Burnkrant & Howard, 1984).
Message Content
Perhaps the most studied message feature is the substantive message content (i.e., the kind of information
included in the message). Among the content variables that have been studied are the quality of the message
arguments, the quantity of information presented, whether the information is focused on emotions or cognitions,
and whether the message includes only content favorable to the advocated side or whether it includes content on
the other side as well.
Argument quality. One of the most manipulated variables in the contemporary literature is the quality or
cogency of the message arguments. As documented earlier in our review, the list of variables that interact with
argument quality in determining persuasion is now quite lengthy. However, despite the large number of studies,
relatively little is known about what makes an argument persuasive. This is because following the initial use of it
for this purpose (Petty et al., 1976), most studies have manipulated argument quality primarily as a
methodological tool to examine whether some other variable increases or decreases message scrutiny, not to
examine the determinants of argument cogency per se (Petty et al., 1993).
32
So, what makes an argument persuasive? Although arguments can take many logical forms (e.g.,
Fogelin, 1982), a typical argument presents some consequence that is likely to occur if the advocacy is adopted.
Based on expectancy value notions (e.g., Fishbein & Ajzen, 1975), consequences that are maximally likely and
desirable should be more compelling than those that are less likely or desirable (Areni & Lutz, 1988; Petty &
Wegener, 1991). Of course, arguments can take other forms. For example, one could argue that a negative
consequence can be avoided if the advocacy were adopted. In such cases, selecting consequences that are
maximally likely and undesirable would be most persuasive (Petty & Wegener, 1991; but see discussion of fear
appeals below). One particularly effect way to convince a person that a consequence is likely is to provide a
causal explanation (Slusher & Anderson, 1996). Thus, a speaker arguing that “instituting comprehensive exams
for college seniors will result in higher paying starting jobs” would be better offer providing an explanation as to
why this result occurs rather than simply citing statistical evidence that it will occur.
Some researchers have argued that arguments will be perceived as better the more they match the way a
person looks at the world. For example, people who think of themselves as religious will find arguments that
appeal to religion to be more persuasive than arguments that are legalistic (Cacioppo, Petty, & Sidera, 1982).
This point of view is generally compatible with functional theories of attitudes (e.g., Smith, Bruner, & White,
1956; Katz, 1960) which hold that individuals and attitude objects can differ in the attributes that are most
important, and cogent arguments would be those that related best to these important attributes (e.g., Shavitt,
1989; Snyder & DeBono, 1989). Thus, attribute importance might be added to desirability and likelihood in
determining argument quality. For example, most students might agree that pretty flowers are very desirable, and
that if tuition were raised, it is very likely that pretty flowers could be planted on campus. However, most
students would probably not agree that having pretty flowers on campus is important. This would attenuate the
cogency of the flowers argument.
An additional factor is the novelty of the consequences proposed. Burnstein and Vinokur (1975; Vinokur
& Burnstein, 1974) have argued that all else being equal, an unfamiliar or unique argument has greater impact
than a familiar one. After all, if a person has already considered an argument previously, it is unlikely to generate
much in the way of new favorable (or unfavorable) responses that could lead to persuasion (or boomerang).
Whether it is necessary to add factors such as importance and novelty to the more traditional likelihood
and desirability dimensions awaits further research. It may turn out that these factors are already considered
when people think about the desirability of consequences. For example, the desirability of flowers may not be
invariant, but may change with the situation. Thus, a student might reason that having pretty flowers is desirable
in a fancy restaurant, but is not desirable on campus if it means raising tuition. If so, then perceived importance
of the consequence might not contribute to overall persuasiveness above and beyond the perceived desirability of
the consequence. Unfortunately, relatively few studies have examined what properties of arguments are the most
critical in mediating persuasive influence. In one exception, Wegener, Petty, and Klein (1994) found that, for
thoughtful individuals at least, happy and sad mood created differences in persuasion by influencing the
perceived likelihoods (but not desirability) of the consequences mentioned in the message (see later discussion of
mood effects).
Argument quantity. Early research suggested that increasing the number of arguments included in a
message enhanced persuasion (e.g., Calder, Insko, & Yandell, 1974; Leventhal & Niles, 1965). Even if people
are not thinking about the arguments, they can reason that the more arguments, the better the position is, or the
more knowledgeable the source is. In fact, when the elaboration likelihood is low, such as when the issue is low
in personal relevance or people have little knowledge on the topic, increasing the number of arguments (Petty &
Cacioppo, 1984a) or making each argument longer (Wood, Kallgren, & Preisler, 1985) increases persuasion
regardless of the quality of those arguments. However, if the elaboration likelihood is high, then increasing the
number of strong arguments enhances persuasion, but increasing the number of weak arguments reduces
persuasion (Petty & Cacioppo, 1984a). Furthermore, under high elaboration conditions, if weak arguments are
added to strong arguments, persuasion can decrease when compared with presentation of the strong arguments
alone (Friedrich, Fetherstonhaugh, Casey, & Gallagher, 1996).
Positive versus negative framing of arguments. As noted previously, arguments can take several forms.
For example, an argument against smoking can be stated in a negative manner such as, “If you don’t stop
smoking, you will die sooner,” or in a positive manner such as “If you stop smoking, you will live longer,”
33
(arguments could also be phrased as failing to take action making positive outcomes unlikely or as taking action
making negative outcomes unlikely, see Petty & Wegener, 1991). Some research has suggested that negatively
framed messages have greater impact on attitudes than comparable positively framed ones (e.g., Meyerowitz &
Chaiken, 1987). What would account for this? If, as noted earlier, negative information gets more weight in
people’s judgments, this might account for the advantage of negative framing. The enhanced impact of
negatively framed arguments is larger when people are motivated to think about each of the pieces of information
presented (Maheswaran & Meyers-Levy, 1990; cf. Rothman, Salovey, Antone, Keough, & Martin, 1993). When
people are not motivated to engage in careful message scrutiny, positively framed messages can be more
impactful than negatively framed ones (Maheswaran & Myers-Levy, 1990). In these low effort situations,
individuals may demonstrate a simple affinity for the more pleasant sounding communication (cf., Zajonc, 1968).
In the health domain, some researchers have also made a distinction between behaviors aimed at
preventing versus detecting disease, with positive (gain) frames being more effective for prevention, but negative
(loss) frames being more effective for detection, behaviors (e.g., Rothman et al., 1993). The prevention versus
detection categorization likely proxies for one or more conceptual variables such as the extent to which the
disease or other positive versus negative aspects of the behavior are salient in that setting. This might be
understood by focusing on the perceived likelihoods of aspects of the message arguments (e.g., Wegener et al.,
1994). Specifically, if thinking about detection behaviors makes the negative aspects of the disease more salient,
then the undesirable outcomes of failing to engage in the detection behavior might seem more likely (i.e., the
impact of negatively framed arguments would be enhanced). If thinking about prevention behaviors makes the
positive aspects of not having the disease more salient, then the desirable outcomes of engaging in the prevention
behavior might seem more likely (i.e., the impact of positively framed arguments would be enhanced; see
Rothman & Salovey, in press; Salovey, Rothman, & Rodin, this volume, for additional discussion of message
framing in health domains).
Individual differences can also be important in determining which framing is more effective. For
example, Higgins’s (1989) self-discrepancy theory notes that individuals who have a discrepancy between their
actual-self (i.e., how one actually is) and their ideal-self (how the person wants to be) are oriented toward
maximizing the presence of positive outcomes and minimizing their absence, whereas those with a discrepancy
between their actual-self and their ought-self (how the person should be) are oriented toward minimizing the
presence of negative outcomes and maximizing their absence. Based on this, Tykocinski, Higgins, and Chaiken
(1995) hypothesized that a negatively framed message would “activate the vulnerability system” and cause
distress for people with an actual-ought discrepancy. This distress was postulated to reduce persuasibility. For
similar reasons, a positively framed message would reduce persuasion for individuals with an actual-ideal
discrepancy. Consistent with this reasoning, actual-ideal discrepancy individuals were more influenced by a
negatively than positively framed message, and actual-ought discrepancy individuals demonstrated the reverse.
However, the mechanism by which distress reduced attitude change was not clear. Perhaps the distress reduced
change by a low effort peripheral process such as classical conditioning. Or, if the message arguments were
strong, perhaps the distress reduced persuasion by reducing processing of the message. Or, perhaps mismatching
the message frame to one’s chronic orientation increased the extent of message processing because a mismatched
message was more surprising.
In an explicit attempt to link message framing to message processing, Smith and Petty (1996)
demonstrated that either positive or negative framing could lead to more processing depending on which type of
frame was expected. Recipients (especially those low in need for cognition) who were led to expect a positively
framed message were more influenced by the quality of the arguments in a negatively framed than a positively
framed message, but recipients who were led to expect a negatively framed message engaged in greater scrutiny
of the arguments in the positively framed communication.
Fear/threat appeals. When very strong negative consequences (e.g., failing an important course, denial of
tenure, death) are implied if an advocacy is not adopted, a threat appeal is being used. These messages are often
referred to as fear appeals since it is assumed that emotional reactions are often induced as well, though
empirically this is not always the case.10 On the surface, at least, it would appear that such appeals would be
very effective since they depict extremely negative consequences as being likely to occur unless the recipient
34
agrees with the message. Thus, in terms of an expectancy-value analysis of argument strength, threat appeals
should be quite cogent. In fact, a meta-analysis of the fear appeals literature indicated that overall, increasing
fear is associated with increased persuasion (Boster & Mongeau, 1984).
Yet, fear appeals are not invariably found to be more effective. One of the earliest studies on fear
appeals suggested the opposite conclusion (Janis & Feshbach, 1953). There are several factors that work against
the effectiveness of fear appeals. First, even if people view the threatened negative consequence as horrific, they
are often motivated by self-protection to minimize the likelihood that some frightening consequence might befall
them (e.g., Ditto, Jemmott, & Darley, 1988). Second, to the extent that the threat is so strong that it becomes
physiologically arousing or distracting, message processing could be disrupted (Baron, Inman, Kao, & Logan,
1992; Jepson & Chaiken, 1990; cf. Baron, Logan, Lilly, Inman, & Brennan, 1994). This would reduce persuasion
if the arguments were strong. Fear is especially likely to reduce message processing if recipients are assured that
the recommendations are effective and the processing might undermine this assurance (Gleicher & Petty, 1992).
It is important to note that when fear reduces message processing, this does not mean that use of fear will be
ineffective in changing attitudes. Rather, it suggests that the effectiveness of some fear appeals could be due to
disrupting processing of weak arguments, or increasing reliance on various peripheral cues (e.g., a reassuring
expert).11
The dominant perspective in this literature is Rogers’ (1983) protection motivation theory. Consistent
with expectancy-value notions, this model holds that fear appeals will be effective to the extent that the message
convinces the recipient that the consequence is severe (i.e., is very undesirable) and very likely to occur if the
recommended action is not followed. Importantly, this theory also holds that effective fear messages should also
convey that the negative consequence can be avoided if the recommended action is followed and that the
recipient has the requisite skills to take the recommended action (see also Beck & Frankel, 1981; Sutton, 1982).
Considerable evidence supports these predictions and has also shown that if people do not believe that they can
cope effectively with the threat, then increasing threat tends to produce a boomerang effect presumably as a
consequence of attempting to restore control or reduce fear (e.g., Mullis & Lippa, 1990; Rippetoe & Rogers,
1987). Interestingly, in young children both self-efficacy and threat produce main effects on evaluations rather
than an interaction as these individuals may not yet have developed a defensive avoidance mechanism (Sturges &
Rogers, 1996) .
Note however, that the protection motivation framework could be applied to virtually any negatively
framed argument. In fact, with rewording, it could also apply to any positively framed argument (i.e., a
promissory appeal will be effective to the extent that it conveys that the consequence is highly desirable, likely,
the person has the skill to bring the positive outcome about, etc.). The emotional reaction of fear in this cognitive
analysis plays relatively little role (see also, Leventhal, 1970; Dillard, in press). However, self-perceived fear
might contribute to persuasion by leading thoughtful recipients to overestimate how bad the consequences are or
how likely they are (Petty & Wegener, 1991; Rogers, 1983).
In sum, threat appeals can be analyzed just as any other message that conveys the likelihood of some
positive or negative consequence occurring. To the extent that the threat appeal also induces an emotional
reaction, however, this emotional reaction presumably can have an additional effect on message acceptance by
serving as a simple cue, biasing message processing, or determining the extent of message scrutiny (see
subsequent discussion of mood effects).
Emotion versus reason in messages. Another issue that was of considerable early empirical interest, and
has re-emerged, is the question of whether emotional (affective) appeals are more or less effective than appeals to
reason and evidence (cognitive appeal). This issue has its roots in the distinction Aristotle drew in his Rhetoric
between “pathos” and “logos” (McGuire, 1969). The initial work on this question either found no overall
difference in effectiveness between the two types of appeals (e.g Knower, 1935) or tended to favor affective over
cognitive messages (e.g, Hartmann, 1936).
Current research suggests that which type of appeal is superior depends on the basis of the attitude under
challenge. In a series of studies, Edwards (1990; Edwards & von Hippel, 1995) concluded that matching was
best. That is, it is better to match the persuasive appeal to the basis of the attitude than to mismatch. For
example, if the attitude is based primarily on emotion, then an emotional appeal is more effective than a cognitive
appeal in changing the attitude. On the other hand, in a separate series of studies, Millar and Millar (1990)
35
concluded that mismatching is best -- if the attitude is based on affect, a cognitive challenge is more effective
than an affective one. These studies used very different methods and materials in testing their hypotheses and
thus it is difficult to pinpoint the reason for the different results obtained. One potentially important difference
between the two sets of studies is that Edwards used attitude objects about which participants had relatively little
information, whereas Millar and Millar used attitude objects for which participants had already established
attitudes (Olson & Zanna, 1993). This could account for the different results because with well formed attitudes,
people may be better able to counterargue a direct attack on the underlying basis of their attitude (Millar &
Millar, 1990). Alternatively, it may be that which effect occurs depends on the cogency of the attack. That is, if
the attack is strong enough to undermine the basis of the attitude, then matching is better, but if the attack is
weak, then mismatching may be superior (Petty, Gleicher, & Baker, 1991). Future work will also likely develop
and rely on more advanced methods for measuring the extent to which existing or experimentally created
attitudes are based on affect or cognition (e.g., Crites, Fabrigar, & Petty, 1994).
One versus two-sided messages. So far, only cases in which the content of a persuasive message presents
just one side of the issue -- the side being advocated -- were considered. In contrast to these typical one-sided
messages, several investigations have examined the consequences of including information on both sides of an
issue. In an influential study, Lord et al. (1979) demonstrated that after examining equally strong evidence on
both sides of an issue, people believed that the evidence on their side was more compelling than the evidence on
the other side (called “biased assimilation”), and they came to believe that their own attitudes toward the issue
had polarized. This is the expected result if one’s attitude biases interpretation of the evidence. Consistent with
this reasoning, Schutte and Fazio (1995) found that the biased assimilation effect was strongest when attitudes
toward the topic were made highly accessible (see also Houston & Fazio, 1989) and recipients were not made
apprehensive about the accuracy of their judgments. The biased assimilation effect is also stronger when people
are emotionally invested in their attitudes (Edwards & Smith, 1996). However, Miller, McHoskey, Bane, and
Dowd (1993) noted that the Lord et al. study and subsequent replications did not demonstrate any actual attitude
change as a result of biased assimilation because only perceived attitude change (polarization) and/or biased
thinking were assessed. To rectify this, they conducted conceptual replications of the Lord et al. research and
included measures of participants’ attitudes toward the topic. Although they replicated the biased assimilation
effect and self-reports of polarization on the issue of capital punishment, after reading essays on both sides of
affirmative action, participants were about equally likely to perceive they had depolarized as polarized. It is
important to note that on neither topic did actual attitude polarization occur. Miller et al. speculated that
participants could have held ambivalent attitudes on these topics which prevented polarization from taking place.
In another relevant study, Pomerantz et al. (1995) found that polarization was only evident among individuals
who were highly committed to their attitudes. Path analyses indicated that increased commitment led to attitude
polarization both directly and as mediated by biased elaboration of the evidence. Finally, Giner-Sorolla and
Chaiken (1996) had people read a message presenting both sides of an issue after assessing their vested interests
on the issue. Consistent with past research, participants engaged in a biased evaluation of the evidence, but
attitude polarization was not found. Rather, attitudes came more in line with participants’ own vested interests
following the message. That is, people who initially held attitudes in conflict with their own interests moved in
the direction of self-interest, but people who already held attitudes in line with their interests tended to
depolarize. Thus, processing a message presenting both sides of an issue does not inevitably result in polarized
attitudes as suggested by the Lord et al. research. Rather, a number of variables related to the strength of one’s
attitude appear to determine whether processing the message will result in polarization, depolarization, or no
change.
In the research using the Lord et al. paradigm, the two-sided message is balanced in that it does not
clearly favor one side over the other. In another line of research, however, two-sided messages are presented in
which one side is clearly the position advocated, whereas the other side is not. Furthermore, the opposition side
is usually presented with weaker arguments than the focal side, or is explicitly refuted. The effectiveness of
these two-sided messages are compared with the effectiveness of a one sided-message in which the opposition
side is not mentioned. The initial research on this suggested that one sided-messages were more effective for
those who initially favored the advocated position, and for those who were relatively low in educational level, but
that two-sided messages were more effective for those in opposition to the position and of higher educational
36
attainment (Hovland et al., 1949).
As reasonable as these early findings appeared, a meta-analytic review of the accumulated literature
failed to provide support (Allen, 1991). Specifically, across the accumulated literature, whether the recipients
agreed or disagreed with the message did not moderate the results. It is possible that the original results were
obtained not because of educational level or prior attitudes per se, but because people who oppose the message
(or are of higher educational levels) are more likely to be aware that opposition arguments exist (and thus dealing
with them directly presumably helps to undermine their implicit impact; Hass & Linder, 1972). Thus, future
meta-analyses and primary research might measure the extent to which recipients are aware of the different sides
of the issue in addition to the recipients’ own attitudes. Allen’s (1991) meta-analysis also discovered that it was
important to distinguish two-sided messages that explicitly refuted the other side (two-sided refutational
message) versus those that presented the other side but did not counter it (two-sided nonrefutational message).
Across the relevant literature, refutational messages were more effective than one-sided communications, but
nonrefutational messages were less effective than one-sided messages. This finding was confirmed in primary
research using a diversity of topics and messages (Allen et al., 1990). In addition, Hale, Mongeau, and Thomas
(1991) found that two sided refutational messages produced more favorable cognitive responses than two sided
nonrefutational messages. Another possible benefit of two sided refutational messages is that they can be more
effective than one-sided communications in instilling resistance to counterattacks (Lumsdaine & Janis, 1953).
To date, only one study has examined the possible multiple roles that two-sided messages can play in
producing attitude change. In this research, Pechmann and Estaban (1994) established three levels of motivation
to process an advertisement using a combination of motivational instructions and situational distraction.
Participants received an advertisement containing strong or weak arguments that favorably compared an
unfamiliar product to a popular product (two sided-message) or did not mention the competing popular brand
(one-sided message). Consistent with the ELM multiple roles notion, under low motivation conditions, argument
quality did not influence purchase intentions but the two-sided ad elicited more favorable reactions than the onesided ad (cue effect). Because the two-sided message favorably compared an unknown brand to a highly liked
brand (positive cue), a simple inference of quality could be responsible for the effect. Under high motivation,
however, only argument strength had an impact. Under moderate motivation, the two-sided (comparative) ad
elicited greater message scrutiny than the one-sided ad. The data further suggested that this enhanced scrutiny
was biased in a favorable direction. Although research is sparse at present, it seems likely that two-sided
messages would generally elicit greater scrutiny than one sided messages (given the need to make comparisons,
resolve discrepancies, etc.) as long as the elaboration likelihood was not constrained by other variables to be high
or low. However, whether two-sided messages serve as positive or negative cues (or biasing agents) will likely
depend on the nature of the two-sided appeal or the message recipients. For example, two-sided appeals can
make the source seem more fair, but might also make the definitiveness of the position seem less clear.
Message Organization
A number of variables have been studied concerning how a message is organized including whether one
should start or end with the strongest arguments (e.g., Sponberg, 1946) and whether the source is presented
before or after the substantive message arguments (e.g., Mills & Harvey, 1972; Pratkanis, Greenwald, Leippe, &
Baumgardner, 1988; see subsequent discussion of the sleeper effect).
Perhaps the most researched variable, however, concerns the placement of competing messages on
different sides of an issue. That is, all else being equal, is it to one’s persuasive advantage to present one’s side
first or second? In the most cited study on this phenomena, Miller and Campbell (1959) relied on learning theory
to predict that primacy would be expected when the two messages are presented together in time but the attitude
assessment is delayed. This is because proactive inhibition would prevent the second message from being
learned as well as the first and the second would have a faster decay. On the other hand, a recency effect would
occur if the second message was presented some time after the first (so that proactive inhibition was minimized)
and the attitude measure was taken shortly after the second message (so there was little decay of the second
message). These predictions were supported in the attitude data, but there was little evidence that memory for the
message arguments was responsible for the effects (see also Insko, 1964). However, if memory for one’s own
favorable thoughts to the communications followed the learning and decay patterns predicted by Miller and
Campbell, the same attitudinal results would be expected. In any case, one moderator of primacy/recency effects
37
is the temporal ordering of the messages and attitude measures.
In reviewing the primacy/recency literature, Haugtvedt and Wegener (1994) noted another possible
moderating factor. Specifically, they noted that several of the studies finding primacy effects tended to use
controversial or familiar issues that likely instilled a high likelihood of elaboration, whereas several of those that
produced recency likely instilled a low elaboration likelihood (see Lana, 1964, for an early review). Using the
ELM, they argued that the greater the processing of the first message, the more likely that a strong attitude would
be formed that would facilitate counterarguing of and resistance to the second message (Haugtvedt & Petty,
1992; Lund, 1925). In support of this hypothesis, Haugtvedt and Wegener (1994) found that when the personal
relevance of the two messages was manipulated to be high, primacy effects were observed, but when the personal
relevance of the two messages was manipulated to be low, recency was observed (see need for cognition section
for additional discussion of effects of elaboration on primacy/recency effects).
Recipient Variables
Recipient variables refer to any aspects that the receiver of the influence attempt brings to the persuasion
situation. That is, the ?recipient” category generally refers to relatively enduring aspects of the individual such as
the person’s demographic category (e.g,. gender, race), or personality and individual skills (e.g., self-esteem,
intelligence). Characteristics of the issue-relevant attitude that the person possesses before the persuasive appeal
is presented are discussed first. These initial attitudes can be determined by a variety of previous experiences, of
course, but are brought to the persuasion situation by the message recipient.
Attitudinal Variables
Influence attempts can be presented to individuals who either have or do not have a prior attitude on the
issue, and if the person has an attitude, this evaluation can be relatively strong or weak. A large literature now
exists on the features that contribute to or are indicative of strong attitudes (Petty & Krosnick, 1995). In general,
strong attitudes are more stable over time, resistant in the face of counterpressure, and have a larger impact on
other judgments and behavior than weak attitudes (Krosnick & Petty, 1995). Attitudes can be strong for a wide
variety of reasons. For example, the attitude can be based on a heritable component (Tesser, 1993), a consistent
and organized belief structure (Chaiken et al., 1995), and so forth. Among the indicators of strong attitudes are
their extremity (Abelson, 1995), accessibility (e.g., Fazio, 1995), and confidence (Gross, Holtz, & Miller, 1995).
Attitude accessibility. What impact do these strength factors have on attitude change? In general, many
of these strength factors would serve in the same roles as other variables. For example, consider the accessibility
of one’s attitude. When the elaboration likelihood is high, then the more accessible the attitude, the more it will
bias message processing in an attitude-consistent direction (Houston & Fazio, 1989). This, of course, would
make movement toward counterattitudinal positions less likely. If the elaboration likelihood is constrained to be
low, however, then the more accessible the attitude, the more likely the person will make snap decisions based on
the salient attitude (i.e., the attitude serves as a peripheral cue; Jamieson & Zanna, 1989). Finally, if the
elaboration likelihood is not constrained, increasing attitude accessibility increases the likelihood that people will
scrutinize the message (Fabrigar, Wegener, Priester, & Petty, in press).
Issue-relevant knowledge. One of the most studied characteristics of attitudes in persuasion contexts is
the extent to which the attitude is associated with high or low amounts of issue-relevant knowledge. Like other
variables, recipient knowledge should be capable of serving in multiple roles. For example, one’s perceived
knowledge could function as a peripheral cue (e.g., I’m the expert so I can reject the disagreeable advocacy”)
especially when the elaboration likelihood is low. Alternatively, knowledge could serve to affect the extent of
information processing through either motivational (“I’ve never heard of that, so I’m curious about it,” or “I’ve
heard so much about that, I’m bored with it”), or ability factors (providing sufficient background to be able to
discern the merits of strong arguments and the flaws in weak ones). Or, knowledge can bias information
processing by motivating or enabling pro or counterarguing depending on whether the message was compatible
or incompatible with one’s existing attitude and knowledge (Petty & Cacioppo, 1986a; Petty et al., 1994).
In one early study, Lewan and Stotland (1961) provided students with factual information about the
unfamiliar country of Andorra or not, and then exposed them to an emotional message attacking the country.
People with some prior knowledge about Andorra were less influenced by the attack than people who had no
prior knowledge. It is not clear, however, if high knowledge recipients resisted because knowledge instilled
greater confidence in their initial opinions, because the knowledge motivated them to process the arguments and
38
the arguments were deemed weak, or because the prior knowledge motivated or enabled counterarguing of the
attack.
More contemporary research has attempted to examine the processes by which knowledge has its impact,
and has consistently demonstrated that people with high amounts of issue-relevant knowledge tend to engage in
greater scrutiny of messages relevant to their knowledge than people with low amounts of issue-relevant
knowledge, and are less reliant on peripheral cues (see Wood, Rhodes, & Biek, 1995, for a review). In addition,
research suggests that people with high knowledge tend to resist influence on counterattitudinal issues more than
people with low knowledge (e.g., Wood et al., 1985) but also tend to be more accepting of proattitudinal
messages (Johnson, Lin, Symons, Campbell, & Ekstein, 1995; Wu & Shaffer, 1987). However, because most
contemporary work on prior knowledge tends to measure knowledge rather than manipulate it, it is not surprising
that high knowledge is associated with more attitude congenial outcomes. This is because knowledge is likely to
be confounded with other variables that would foster this bias (e.g., high levels of confidence, attitude extremity,
etc.). If knowledge was not confounded in this way, it would presumably be more likely to determine the extent
of information processing activity with bias resulting from other motivational and ability factors (e.g.,
confidence, extremity). This is conceptually similar to our analysis of research on measured attitude importance
or involvement. Consistent with this view, Biek, Wood, and Chaiken (1996) found that people with high
measured knowledge engaged in greater biased processing of information only when their attitudes were
associated with considerable affect. When knowledge was high but affect was low, processing was more
objective.
Demographic Variables
Gender. Early research tended to show that women were more susceptible to influence than were men
(e.g., Janis & Field, 1959; Knower, 1936). One explanation was based on culture. That is, society had greater
conformity expectations for women (Hovland & Janis, 1959). Another explanation relied on the presumed
greater message reception skills of women (McGuire, 1969). Early research also suggested that sometimes this
effect was due to the gender of the influence agent -- when this was controlled, the gender difference disappeared
(Weitzenhoffer & Weitzenhoffer, 1958). McGuire (1968) also speculated that the effect might be due to the fact
that the experimental materials in most studies of the time were constructed by men.
More contemporary analysts have concluded that a small gender difference exists and is exacerbated
when the study is conducted by a male investigator or influence is assessed in a group-pressure situation (Cooper,
1979). For social reasons, males might be more effective in eliciting compliance from females than are other
females, and women might be more interested in social harmony than are men (due in part to their early
socialization or their expected social roles; Eagly, 1978; Eagly & Wood, 1991). For example, in some situations,
if one’s gender is salient and motivation or ability to carefully scrutinize the merits of the issue are reduced, a
cultural norm might determine the extent of influence (e.g., “as a man, I shouldn’t give in”).
These explanations suggest that gender differences will persist as long as cultural factors remain similar.
Cultural factors do not account for all of the variance in gender effects, however. Some of the effect can be
attributed to the nature of the influence topic. For example, in some cases, the topic of the message might
determine how much scrutiny is given to the message (e.g., “this message is relevant to all women, so I should
think about it”). Or, if the topic is one on which men have stronger attitudes than women (e.g., due to greater
attitude-supportive knowledge), then women might be more influenceable since they would be less motivated or
able to defend their attitudes. If the attitude strength differences are reversed, however, then men would be more
influenceable (Cacioppo & Petty, 1980; Sistrunk & McDavid, 1971). In addition, the message content can be
critical even if the topic is held constant. For example, different arguments might appear stronger (on average) to
each gender. One study, for instance, found that women were more susceptible to an appeal to sympathy than
reciprocity (Fink, Rey, Johnson, Spenner, Morton, & Flores, 1975).
Age. Effects of age on influenceability have been of intense interest in the 1990s due to charges that
prosecutors had subtly (and not so subtly) influenced young children to testify to events that did not occur in
some high profile child molestation cases. Early research examined susceptibility to influence by exposing
individuals of different ages to various suggestion and hypnotism tests. This research suggested that young
children were quite open to suggestion (e.g., Messerschmidt, 1933), and more recent studies reinforce this
conclusion (e.g., Ceci & Bruck, 1993). Another approach to examining age differences has been to compare the
39
test-retest attitude correlations for individuals of different ages. In general, these studies have indicated greater
stability in attitudes for older than younger individuals (e.g., Alwin, Cohen, & Newcomb, 1991).
A number of hypotheses have been put forth regarding the empirical relation between age and
susceptibility to influence. Some investigators have argued that people generally become less susceptible to
influence as they grow older (e.g., Glenn, 1980). Others have argued that the decrease in susceptibility is not
gradual as one ages, but is rather abrupt as individuals leave their “impressionable” childhood and young adult
years behind them (Mannheim, 1952). Still others have suggested a curvilinear relationship with younger and
older individuals being most susceptible to change (e.g., Sears, 1981), though some evidence suggests that the
increased susceptibility among the oldest individuals could be due to increased measurement error (Krosnick &
Alwin, 1989).
Because of the general consensus that young people, at least, are more susceptible to persuasion than
older adults (Visser & Krosnick, 1996), Sears (1986) argued that the typical laboratory study with college
students overestimates the ease of attitude change in the general population. Of course, the goal of laboratory
research is not typically to provide population estimates, but rather to examine some conceptual hypothesis.
Furthermore, it seems unlikely that age per se relates to influenceability. Rather, a number of factors associated
with age are probably responsible for any link between aging and attitude change and these factors could be
studied within the population of college students (Petty & Cacioppo, 1996). For example, young people may
appear to exhibit less stability in attitude surveys because they happen to be exposed to more challenges to their
attitudes than are older individuals (Tyler & Schuller, 1991). Alternatively, as people grow older, their
knowledge on many issues increases. As noted previously, knowledge can help a person resist an incoming
message. In addition, some people might hold beliefs (such as “older is wiser”) that confer resistance even
though other people of the same age do not hold such beliefs. In addition, individuals of different ages will likely
find different topics of greater or lesser interest and different arguments as being of higher or lower quality. It is
likely that each of these factors operates and accounts, at least in part, for observed age differences in different
situations. Importantly, each of these conceptual questions can be studied within any given sample of
individuals.
Personality/Skills
Although individual differences in the propensity to engage in evaluation and form attitudes have
recently been uncovered (Jarvis & Petty, 1996), the search for a personality variable that captures a general
susceptibility to persuasion across a wide variety of situations has met with little success (e.g. Hovland & Janis,
1959; see Eagly, 1981). Some early evidence indicated that individuals’ responses to various suggestibility tests
showed a small positive correlation (see Eysenck & Furneaux, 1945; Hilgard, 1965) as did responses to various
conformity situations (e.g., Abelson & Lesser, 1959). More success in persuasion situations has come from
research examining specific individual differences in skills and personality traits. In fact, a wide number of
specific traits have been linked to persuasion outcomes including: uncertainty orientation (Sorrentino, Bobocel,
Gitta, Olson, & Hewitt, 1988), internal-external locus of control (e.g., Sherman, 1973), public self-consciousness
(Carver & Scheier, 1982), need for closure (Webster & Kruglanski, 1994), and many others. Our focus is on the
stable traits that have achieved the most research attention.
Intelligence. Primary studies have provided support for both positive (e.g., Cooper & Dinerman, 1951)
and negative (see Crutchfield, 1955) relations between intelligence and influenceability. As noted previously,
McGuire (1968) argued that which effect occurred should depend on whether reception or yielding processes
were more important. For example, with a complex but cogent message, reception would be more important and
thus intelligence would be positively related to persuasion. With a simple message, yielding would be more
important, and intelligence would be negatively related to persuasion because highly intelligent people would be
more resistant (see also Eagly & Warren, 1976). That is, intelligent individuals would likely have greater issuerelevant knowledge on many issues and thus have a greater ability to defend their current positions. However,
their greater knowledge might also enable them to see the merits in complex arguments that would pass by less
intelligent people.
A meta-analytic examination of the accumulated literature on intelligence and persuasion revealed that
increased intelligence was generally associated with decreased persuasion (Rhodes & Wood, 1992). Given that
intelligent people have a greater ability to scrutinize messages than people of less intelligence, this finding
40
probably implies that most counterattitudinal messages used in experiments are not so cogent that intelligent
people cannot counterargue them when they are motivated to think about them. Although it has not been studied
explicitly, perceived intelligence could also serve as a peripheral cue when motivation to think is low. For
example, someone might reason that “I’m probably more intelligent than the source, so why should I change my
view?” This inference might be less likely if the message appeared to be very complex, thus making the source
appear more competent.
Self-esteem. The overall regard that a person has for him or herself -- self-esteem -- has also been
subjected to McGuire’s reception/yielding analysis. As with intelligence, some research has demonstrated a
positive relationship with persuasion (e.g., Berkowitz & Lundy, 1957) whereas other research has demonstrated a
negative relationship (e.g., Janis, 1954). McGuire (1968) thus suggested that the relation between self-esteem
and persuasion should be positive when reception processes dominate, but negative when yielding processes
dominate (see also Nisbett & Gordon, 1967). If both processes operate simultaneously, then one would expect a
curvilinear relationship between self-esteem and persuasion. Although a meta-analysis of the literature revealed
that this curvilinear relationship holds (Rhodes & Wood, 1992), it is not entirely clear that the reception/yielding
model provides the best account of these data.
For example, Skolnick and Heslin (1971) examined previously published studies on self-esteem and
persuasion and had the messages used in these studies rated for their overall persuasiveness and
comprehensibility. Comprehensibility (a stand-in for reception) did not account for the relation between selfesteem and persuasion, but the rated quality of the arguments did. Specifically, when Skolnick and Heslin
divided the accumulated studies into those finding a positive relationship between self-esteem and persuasion and
those finding a negative relationship, they found that the positive relationship studies used more cogent
arguments than the negative relationship studies. This finding is consistent with the view that argument quality
was more important in determining the attitudes of high than low self-esteem individuals. Low self-esteem
individuals might have little need to scrutinize the merits of a communication because they would believe that
most people are more competent than they are and thus, the message can be accepted on faith. A high selfesteem person, however, would have the confidence to scrutinize the message. If situational factors inhibited
message processing (e.g., high levels of distraction), then one might expect high self-esteem individuals to be
generally more resistant than low self-esteem individuals because they would be more likely to reason that their
own opinion was as good or better than that of the source. High self-esteem individuals would presumably be
more susceptible to the “ownness bias” (Perloff & Brock, 1980) that was described previously.
Self-monitoring. Snyder (1974) introduced the notion that some people -- called high self-monitors -- are
very sensitive to cues that indicate socially appropriate behavior in a given situation, whereas other people -called low self-monitors -- are more guided by their internal beliefs and values. High and low self-monitors have
differed in a number of important ways. For example, because internal beliefs are more important to low selfmonitors, they are more susceptible to dissonance effects (Snyder & Tanke, 1976). Most research on selfmonitoring has examined the notion that attitudes serve different functions (see Katz, 1960; Smith et al., 1956)
for people who are high versus low in self-monitoring, and each group should be more persuaded by messages
that matched (versus mismatched) the function served by their attitudes. More specifically, attitudes should serve
a social-adjustive function for high self-monitors and thus they should be especially influenced by arguments that
make claims about the social images one can attain by agreeing with the advocacy. In contrast, the attitudes of
low self-monitors should serve a value-expressive function and thus these individuals should be especially
influenced by arguments that make claims about the underlying merits or true qualities of the issue or object
under consideration. Several tests of these ideas proved congenial. That is, high self-monitors were more
influenced by appeals to image or status whereas low self-monitors were more influenced by messages that made
appeals to values or quality (e.g., DeBono, 1987; Snyder & DeBono, 1989).
In contrast to the functional hypothesis that focuses on the overall effectiveness of messages that match
an attitude’s function, Petty and Wegener (1996) hypothesized that people would engage in greater scrutiny of
message content that matched the functional basis of their attitudes. This implies that when the message
arguments are strong, functional matching will lead to enhanced persuasion (as postulated by the functional
theorists), but when the message arguments are weak, functional matching can lead to reduced persuasion. Some
suggestive evidence for this was provided by DeBono and Harnish (1988). This study found that high and low
41
self-monitors engaged in greater scrutiny of a message when the message source matched the functional basis of
the attitudes. That is, high self-monitors engaged in greater scrutiny of the arguments when they were presented
by an attractive source (who might be expected to make an image appeal) than an expert source (who presumably
would make a quality appeal), whereas low self-monitors demonstrated the reverse pattern. In this study, the
actual content of the communications was the same. Thus, observed processing effects could be due to the
messages not entirely matching the expectations associated with the sources of the messages. Also, in other
settings, message sources that serve the needs of message recipients have led to less rather than more processing
of the message content (e.g., DeBono & Klein, 1993). In order to examine whether individuals would engage in
greater processing of messages when the substantive content of the message actually matched the function served
by the attitude, Petty and Wegener (1996) had high and low self-monitors read image or quality appeals that
contained strong or weak arguments. Consistent with the processing view, the cogency of the arguments had a
larger effect on attitudes when the message matched rather than mismatched the functional basis of the attitude.
According to the ELM, functional matching should influence the extent of message scrutiny primarily
when other factors in the persuasion setting have not already established the elaboration likelihood to be very
high or very low. If circumstances constrained the overall likelihood of elaboration as very low, a functional
match could serve as a simple cue to enhance persuasion. For example, if a source simply asserted that his or her
arguments were relevant to a person’s values, a low self-monitor might be more inclined to agree than a high selfmonitor by reasoning, “if it speaks to my values, it must be good.” On the other hand, if circumstances rendered
the likelihood of elaboration as high, matching the content of the message to the functional basis of the attitude
might bias processing. For example a high self-monitor would be more motivated to generate favorable thoughts
to a message that made an appeal to image rather than an appeal to values.
Need for cognition. In deference to Cohen’s (1957) pioneering work on scaling individual differences in
cognitive motivation, Cacioppo and Petty (1982) called their scale measuring individual differences in the
motivation to think, the need for cognition scale. Individuals who are high in need for cognition enjoy cognitive
activities and engage in them when given the chance. Individuals who are low in need for cognition are cognitive
misers who avoid effortful thinking unless situational demands require it (see Cacioppo, Petty, Feinstein, &
Jarvis, 1996, for a review).
Research has supported the ELM notion that people who are inclined to engage in effortful cognitive
activity are more influenced by the substantive arguments in a persuasive message (e.g., Cacioppo et al., 1983)
and are less influenced by simple peripheral cues (e.g., Haugtvedt et al., 1992) than are those who are less
inclined to think. It is important to note two additional findings. First, the more extensive thinking of individuals
high in need for cognition is not necessarily objective. In fact, two studies have provided evidence that moods
can introduce a significant bias to the thought content of people high in need for cognition (Petty, Schumann,
Richman, & Strathman, 1993; Wegener et al., 1994). Second, research indicates that even low need for cognition
individuals can be motivated to scrutinize the message arguments and eschew reliance on cues if situational
circumstances are motivating -- such as when the message is of high personal relevance (Axsom et al., 1987), the
source is potentially untrustworthy (Priester & Petty, 1995), or the message content is surprising (Smith & Petty,
1996).
Two empirical discrepancies have arisen regarding need for cognition and attitude change. The first
involves the mere thought effect (in which thinking about one’s attitude leads to a polarization of that attitude;
see Tesser, 1978). Given the greater propensity of high need for cognition individuals to engage in thought, one
might expect them to show greater attitude polarization following a period of reflection on their attitudes.
Although one study supported this idea (Smith, Haugtvedt, & Petty, 1994), another found the opposite (Leone &
Ensley, 1986). Research by Lassiter, Apple, and Leach (1994) suggests that this discrepancy might be resolved
by considering the different instructions used in the two studies. Specifically, Lassiter et al. found that when
participants were instructed to think about their attitudes (as in Leone & Ensley), low need for cognition
individuals showed greater polarization than high need for cognition individuals. However, when no explicit
instructions to think were provided (as in Smith et al., 1994), high need for cognition individuals showed greater
polarization. This suggests that when thinking is instructed rather than spontaneous, high need for cognition
individuals may consider all sides of the issue and thus show moderation rather than polarization.
The second discrepancy involves primacy and recency effects. Specifically, one study found that
42
individuals high in need for cognition demonstrated greater primacy in judgment than individuals low in need for
cognition (Kassin, Reddy, & Tulloch, 1990). This result is consistent with the view presented earlier that high
amounts of thinking about early information can enhance counterarguing and rejection of later information
(Haugtvedt & Petty, 1992; Haugtvedt & Wegener, 1994). However, another study found that low need for
cognition individuals demonstrated greater primacy in judgments than individuals high in need for cognition
(Ahlering & Parker, 1989). This result is consistent with the view that low amounts of thinking can cause
individuals to freeze on the early information and ignore subsequent information (Kruglanski & Freund, 1983).
Petty and Jarvis (1996) noted that this discrepancy might be resolved by considering the fact that in the studies
finding greater primacy under high thinking conditions, the materials presented two clear sides to an issue (e.g.,
the prosecution and defense positions in a trial; Kassin et al., 1990; independent pro and con messages on an
issue; Haugtvedt & Wegener, 1994), and participants first received information from one source on one side of
the issue and then the other source on the other side. In contrast, studies finding greater primacy under low
thinking conditions have not divided the information neatly into two sides. Rather the information came in a
continuous stream (Ahlering & Parker, 1989). In the former procedure, thoughtful individuals would likely form
a thoughtful opinion after the initial side and then be biased in processing the second side. When there are no
clear sides, however, highly thoughtful individuals would be more likely to process all of the information prior to
rendering a judgment. This would attenuate any primacy effect.
Context Variables
Context refers to any factors related to the setting in which the communication is presented. This is a
broad category of variables that includes such features as any distractions that are present in the setting, whether
the message is repeated or not, whether the surroundings create a pleasant or unpleasant atmosphere, whether
people are forewarned of the message content or not, expect to discuss the issue with others, and so forth. The
impact of these variables on the processes of persuasion are discussed next.
Distraction
It is not uncommon for people to simultaneously encounter a persuasive message and engage in one or
more other tasks, or have various kinds of distracting stimuli present that tax cognitive capacity. Festinger and
Maccoby (1964) proposed that distraction could increase persuasion by interfering with counterarguing. Early
research supported this proposition by showing that increases in distraction resulted in increases in persuasion
when attention was focused on the message, but not on the distracting events (e.g., Insko et al., 1974) and that
increases in distraction led to decreases in counterarguments measured in thought-listings (e.g., Osterhouse &
Brock, 1970; see Baron, Baron, & Miller, 1973, for a dissonance interpretation). Petty et al. (1976) tested a more
general thought-disruption view of distraction by crossing distraction with a manipulation of argument quality.
According to the disruption of processing view, distraction should lead to decreases in whatever cognitions
would normally have occurred. Thus, if the dominant cognitive responses would have been favorable when no
distraction was present -- as would be the case if the arguments were strong -- then distraction should disrupt
these favorable thoughts and reduce persuasion. However, if the dominant cognitive responses would have been
unfavorable when no distraction was present -- as would be the case if the arguments were weak -- then
distraction should disrupt these unfavorable thoughts and increase persuasion. The data supported this
hypothesis and provided evidence for the view that distraction influences attitudes by disrupting one’s thoughts.
At the same time, this study provided strong support for the cognitive response model of persuasion.
Further support for an interpretation of distraction effects as due to disruption of ability to process the
content of message arguments comes from studies in which manipulations of distraction have been crossed with
manipulations of persuasion cues. For example, Kiesler and Mathog (1968) found that a high-credibility source
led to more favorable attitudes than a low-credibility source to a greater extent when distraction was high rather
than low (see also Miller & Baron, 1968, cited in Baron et al., 1973). One ironic effect of distraction is that it
can lead people to favor positions they intend to disfavor. Specifically, in one study, students were asked to try
not to believe in the conclusion of a message. When not distracted, those trying to disbelieve had no problem
doing this. When distracted, however, the opposite resulted -- those attempting not to believe came to favor the
proposal more presumably because distraction prevented them from suppressing the unwanted belief (see
Wegner, 1994, for discussion of this and other examples of thought suppression).
Audience Reactions
43
When a persuasive message is encountered, other recipients of the message might provide noticeable
reactions to the message (i.e., of agreement or disagreement with the message). For example, Axsom et al.
(1987) presented research participants with an audiotaped strong or weak message accompanied by taped
audience reactions. The message advocacy (probation as an alternative to imprisonment) was introduced as
being high in personal relevance (i.e., considered for the participants’ own state) or low in personal relevance
(i.e., considered for a distant state). In addition, research participants were classified as either high or low in
need for cognition (Cacioppo & Petty, 1982). When elaboration likelihood was lowest (i.e., when personal
relevance and need for cognition were both low), audience agreement reactions (i.e., taped applause that was
consistent and enthusiastic) led to more favorable attitudes than audience disagreement (i.e., inconsistent and
sparse applause). In contrast, when elaboration likelihood was high (i.e., when personal relevance, need for
cognition, or both were high), audience reactions had no effect on favorability of attitudes. Instead, under such
conditions, only message quality influenced attitudes. That is, when elaboration likelihood was low, audience
applause acted as a persuasion cue, but when elaboration likelihood was high, the audience cue had little effect.
In this study, unambiguous strong or weak messages were used. If ambiguous messages were used,
however, one might find evidence of biases in processing under high elaboration conditions (perhaps looking for
reasons why other message recipients seem to agree with the ambiguous arguments). In one study relevant to
potential biased processing instigated by audience reactions under high elaboration conditions, Petty and Brock
(1976) investigated the effects of hecklers on persuasion. They found that when a speaker ignored hecklers or
provided irrelevant responses, audience members expressed less agreement with the message than when no
heckling occurred (perhaps because of negatively biased processing when others voiced their counterarguments
to the message). When the source provided relevant responses to the hecklers (i.e., counterarguing the
counterarguments), however, this reduced the deleterious effects of heckling.
Of course, other reactions of audiences could be studied, some of which might be likely to influence the
amount of message processing if other factors do not constrain processing to be high or low. For example, one
could imagine audience reactions that communicate to other message recipients that the message is interesting or
important versus uninteresting or unimportant. If so, such reactions might be especially likely to affect how
much other audience members process the message.
Forewarning
At times, people receive persuasive messages when they have already learned either the position to be
advocated by the message or the intent of the provider(s) of the message to persuade. Some studies of
forewarning have included both types of forewarning (e.g., Brock, 1967) though, as noted by Papageorgis (1968),
the two types are conceptually distinct. In fact, most research has independently investigated the effects of each
type of forewarning.
Forewarning of content. McGuire and Papageorgis (1962) noted that warning of the content of a
persuasive message might motivate people to consider information that supports their current opinion on the issue
and counterargue opposing positions in anticipation of the message. If this is the case, one should find that
whether or not there is time between the forewarning and the communication creates a difference in the
resistance to the message (with time for preemptive counterarguing leading to greater resistance). Consistent
with such a possibility, Freedman and Sears (1965) found that forewarning teenage students of a message
opposing teenage driving led to little resistance to the message if the forewarning was provided immediately
before the message but to significantly more resistance if provided 10 minutes prior to the message (see also Hass
& Grady, 1975).
Petty and Cacioppo (1977) found additional support for the anticipatory counterargument position by
showing that thought listings revealed significantly more incidence of anticipatory counterargumentation when
people were forewarned of counterattitudinal message content than when they listed thoughts but were not
forewarned, and by showing increased physiological activity indicative of concentrated negative thought during
the postwarning-premessage period (Cacioppo & Petty, 1979a). In addition, Petty and Cacioppo (1977) noted
that it was not the forewarning per se that induced resistance to an attacking appeal, but rather the activation and
consideration of attitude-relevant knowledge. That is, activation of attitude-relevant knowledge (accomplished
by asking people to list thoughts on the topic without forewarning them of the content of the upcoming message)
prior to receipt of an attacking message was sufficient to produce resistance to the appeal equal to that of
44
forewarning.
Consistent with the idea that activation of attitude-relevant knowledge was responsible for resistance
when people are forewarned of persuasive content, it has been shown that resistance only occurs to the extent that
the message is on a topic that is personally important or involving to the message recipient (Apsler & Sears,
1968). Attesting to the importance of motivational and ability factors, Chen, Reardon, Rea, and Moore (1992)
found that a forewarning of message content on a counterattitudinal issue led people to resist the message and
generate unfavorable thoughts to both strong and weak arguments primarily when the issue was personally
involving and they were not distracted (see also Freedman, Sears, & O’Connor, 1964; Romero, Agnew, & Insko,
1996).
Forewarning of persuasive intent. If a person does not even know what a communication is about, it
seems unlikely that he or she could generate anticipatory counterarguments. Therefore, one must look to other
mechanisms to account for the effects of forewarning of persuasive intent (see Cialdini & Petty, 1981). An
empirical indication that forewarning of persuasive intent works differently than forewarning of persuasive
content is that forewarning of intent is equally effective, regardless of the amount of time between the
forewarning and the message (e.g., Hass & Grady, 1975). It could be that intent to persuade is regarded as an
intended restriction of freedom to think or act in some way. If so, a recipient of such a forewarning might
experience reactance (Brehm, 1966) that would instigate counterarguing once the persuasive appeal has begun
(counterarguing cannot begin until the topic of the appeal is known; see Fukada, 1986, for differential self-reports
of reactance by warned and unwarned message recipients). In order to investigate whether counterarguing during
the message was responsible for the effects of forewarning of persuasive intent, Petty and Cacioppo (1979a)
exposed research participants to a message arguing that senior comprehensive exams be instituted after the issue
was described as either high or low in personal relevance and participants were either forewarned of persuasive
intent or not. Results showed that forewarning led to less favorable opinions of the advocacy, but only
significantly so when the topic was high in personal relevance. This suggested that forewarning was not acting as
a simple rejection cue (which would have worked under low, rather than high, relevance), but was acting to bias
the processing that occurred when personal relevance was high. That is, relevance increased processing and led
to greater acceptance of the strong arguments in the message when no forewarning was present, but when a
forewarning was present, then the relevance-inspired processing became biased.
Additional evidence for the view that forewarning of persuasive intent biases processing of a message
comes from research demonstrating that such warnings are effective in reducing persuasion when they are
presented before rather than after the message (Kiesler & Kiesler, 1964) and produce resistance only when
message recipients are not distracted during message presentation (Watts & Holt, 1979).
Anticipated Discussion or Interaction
A number of studies have examined how opinions are modified when people are asked to interact with or
be accountable to some other person or persons. Like other variables, anticipation of discussion or accountability
to others has produced a number of effects. In general, when people feel they are accountable or might have to
justify their stance to others, this engages concerns about one’s social appearance (Cialdini et al., 1976; Leippe &
Elkin, 1987; Tetlock, 1992). How one goes about maximizing a favorable impression, however, varies with the
situation. Consider first the case in which a person expects to discuss or be held accountable for an issue that is
not particularly important or is somewhat unfamiliar. In such cases, if the opinions of the audience (or person to
whom one is accountable) are unknown, people tend to diligently think about any information presented on the
issue in an attempt to adopt the best or most justifiable position (Chaiken, 1980; Petty et al., 1980; Tetlock,
1983). If no information is presented, however, and the audience’s opinion is unknown, people tend to adopt a
moderate and presumably defensible position in anticipation of interaction with the audience (Cialdini et al.,
1976). When the audience’s opinions are known, communications to these audiences tend to be biased in favor
of the audience's opinions and people tend to shift their own opinions toward those expected to be held by the
audience (e.g., Chen et al., 1996; Higgins & McCann, 1984; Tetlock, 1983).
When the issue is important or people have a previous commitment to a particular position, things
change. Specifically, being held accountable or expecting to discuss the issue leads people to justify their initial
positions (Lambert et al., 1996; Tetlock, Skitka, & Boettger, 1989) which can result in attitude polarization rather
than moderation (Cialdini et al., 1976; see Cialdini & Petty, 1981; Tetlock, 1992, for additional discussion).
45
Channel (Message Modality)
Any given persuasive message can be presented in various ways, through various media. Although the
messages presented in social psychological studies are often written (on paper or computer screen), audiotaped,
or videotaped, one conceptual variable likely represented in these categories is a distinction between self-paced
and externally-paced presentation. That is, a written message is typically self-paced (i.e., a reader can go back to
read and reread anything that he or she likes, e.g., Chaiken & Eagly, 1976; for an exception, see Mackie &
Worth, 1989). When a communication is audiotaped or videotaped, however, it is more likely to be controlled by
someone other than the message recipient (e.g., Andreoli & Worchel, 1978; Chaiken & Eagly, 1976).
What persuasive effects are such differences likely to have? Although little basic research has
investigated differences in communication modality, most of the relevant work suggests that some methods of
presenting a message receive greater scrutiny than others. As might be expected, self-paced messages generally
receive greater scrutiny than externally paced messages (perhaps because it is more difficult to thoroughly
scrutinize the content of an externally-paced message, at least when it is reasonably complex in nature).
Consistent with this idea, when Eagly and Chaiken (1976) presented research participants with a complex, cogent
message, they found that the message led to greater persuasion and recall of message arguments when it was
written (self-paced) than when the message was audiotaped or videotaped (externally-paced). This is to be
expected assuming that the message arguments were strong.
In addition, it has been shown that peripheral persuasion cues such as communicator credibility or
likeableness tend to have a greater impact when the message is videotaped (i.e., externally-paced) rather than
written (self-paced; e.g., Andreoli & Worchel, 1978; Chaiken & Eagly, 1983). These effects could be due to the
difficulty of processing complex material when externally paced (as mentioned above), but could also be due in
part to the salience of source cues across the different versions of the message. Many visual aspects of the person
or setting in which a source speaks could enhance perceptions of credibility or likeableness, and these features of
the person and setting can only truly affect message recipients when the person and setting are clearly pictured
with the message (which can occur when a message is videotaped, but not audiotaped). Finally, the modality of
the message itself could serve as a cue or a determinant of biased processing in that some people might, for
example, be more impressed with messages appearing in the print media (e.g., newspapers, books, magazines)
than on television.
Mood
A variety of events or aspects of one’s environment can change the way one feels. In empirical
investigations of effects of mood on persuasion, this has generally been accomplished by providing people with
written, audiotaped, or videotaped material that is generally very pleasant or unpleasant prior to message
exposure, thus setting a persuasion context in which the recipient’s mood is positive, negative, or neutral. Effects
of positive and negative moods have been examined across the elaboration continuum.
According to the ELM, mood can serve in the same multiple roles as other variables (Petty, Gleicher, &
Baker, 1991). Thus, when the likelihood of issue-relevant thinking is low, a person's mood should impact
attitudes by a peripheral process. Consistent with this view, early investigations of mood and persuasion were
often guided by classical conditioning notions of a direct association between the attitude object and the person's
affective state (e.g., Griffit, 1970; Zanna et al., 1970). More recently, affective states have been postulated to
influence attitudes by a simple inference process in which misattribution of the cause of the mood state to the
persuasive message or to the attitude object occurs (e.g., I must feel good because I like or agree with the
message advocacy; see Petty & Cacioppo, 1983; Schwarz, 1990). Importantly, these direct effects of mood on
attitude seem to be more likely when elaboration likelihood is low than high (e.g., Gorn, 1982; Petty et al., 1993).
As the likelihood of elaboration increases, mood takes on different roles. Specifically, when the
elaboration likelihood is more moderate, mood has been shown to have an impact on the extent of argument
elaboration. Competing theoretical positions have been put forward to explain effects of mood on message
processing. The "cognitive capacity" and “feelings-as-information” views both predict that happy moods disrupt
processing of message content (see Mackie & Worth, 1991; Schwarz, Bless, & Bohner, 1991). For the cognitive
capacity view, this is because happy moods activate positive thoughts in memory which occupy a person's
attentional capacity and render the message recipient less able to process incoming information (Mackie &
Worth, 1989, 1991). In comparison, the feelings-as-information view states that negative moods signal that
46
something is wrong in the environment and that some action is necessary, whereas positive moods indicate that
no scrutiny of the environment is required (Schwarz, 1990). Because of this, negative states generally instigate
active processing strategies in order to deal with problems in the environment, but positive states do not
(although active processing can be done in positive states if other goals that necessitate such processing become
salient; see Bless, Bohner, Schwarz, & Strack, 1990). Although these views have been used to account for a
large number of studies, virtually all of the persuasion experiments investigating the effects of mood on
processing have used counterattitudinal or depressing messages (e.g., on topics such as acid rain, gun control,
student tuition increases, etc.; see Wegener & Petty, 1996, for a review). Because of this, one might also account
for deficits in message processing by happy people by considering the possibility that happy people pay more
attention to the hedonic consequences of their actions than people in neutral or sad states. If this were the case,
then happy people might be especially avoidant of activities perceived as likely to be depressing, but might
especially engage in uplifting activities. Wegener and Petty (1994) developed this hedonic contingency idea by
noting that hedonic rewards (i.e., feeling better rather than worse after engaging in an activity) are more
contingent on the scrutiny of the hedonic consequences of action in happy than in sad states.
Empirical tests of the hedonic contingency possibility have been encouraging. Wegener, Petty, and Smith
(1995, Experiment 1) found that use of a proattitudinal message led to greater message processing from happy
than neutral people (which would not be predicted by either the “cognitive capacity” or “feelings-as-information”
views). Wegener et al. (1995; Experiment 2) manipulated the introduction of messages so that the same
arguments could be used to support either a proattitudinal (uplifting) position or a counterattitudinal (depressing)
position, and found that happy people processed the arguments more when they addressed an uplifting proposal
than when they addressed a depressing proposal, whereas sad people processed the messages to the same extent,
regardless of the uplifting or depressing introduction. Organized another way, this result showed that the hedonic
contingency idea was capable of accounting for the past effects of mood on message processing. That is, when a
depressing version of the topic was used (as in the past research, e.g., Bless et al., 1990; Mackie & Worth, 1989),
happy people processed the messages less than sad people (replicating the past result). However, when an
uplifting version of the topic was used, the same arguments were processed more by happy people than by sad
people.
Thus, it appears that happy people do not universally process information less than neutral or sad people.
Rather, happy people engage in cognitive tasks to the extent that the task is viewed as enabling the person to
remain happy (or keep from feeling badly). This view might account for why happy moods have been found to
enhance cognitive activities in some areas outside the persuasion domain (e.g., enhancement of creative problem
solving, Isen, Daubman, & Nowicki, 1987; enhancement of generation of similarities and differences between
pairs of targets, Murray, Sujan, Hirt, & Sujan, 1990). Of course, within the multiple roles framework, such
effects on amount of cognitive processing should be most likely if background factors do not constrain
elaboration likelihood to be extremely high or low.
When the elaboration likelihood is high and people are processing the message arguments already, the
ELM holds that affective states can influence attitudes by influencing the nature of the thoughts that come to
mind. Positive mood can facilitate the retrieval of positive or inhibit the retrieval of negative material from
memory (e.g., see Blaney, 1986, Bower, 1981). Thus, a person's mood during message processing can be related
to the favorability of the cognitive responses generated. Support for this possibility has been found in a number
of studies (e.g., Breckler & Wiggins, 1991; Mathur & Chattopadhyay, 1991), some of which explicitly varied the
elaboration likelihood. In two studies, Petty et al. (1993) found that positive mood had an impact on the
favorability of thoughts of subjects high in elaboration likelihood (i.e., for people high in need for cognition or
under high message relevance conditions), and these thoughts influenced attitudes. In contrast, when elaboration
likelihood was low, mood did not influence thought valence, but rather directly influenced attitudes presumably
by a low effort peripheral process.
This is not to say that positive moods should always bias processing toward being more favorable toward
the position advocated or that negative moods should invariably render the message conclusion less acceptable.
For example, negative moods have been shown to make negative events seem more likely than positive events
(e.g., Johnson & Tversky, 1983) and positive events and behaviors are seen as more likely in positive as opposed
to negative moods (e.g., Mayer, Gaschke, Braverman, & Evans, 1992). Thus, to the extent that message
47
arguments include statements that a plan should be followed in order to avoid negative consequences, negative
moods could actually lead high elaboration recipients to view these arguments as more compelling (see Petty &
Wegener, 1991; Wegener et al., 1994).
Repetition of the Message
Models of persuasion that viewed attitude change as acquisition of a new verbal habit (e.g., Hovland et
al., 1953) suggested that repetition of a message would enhance persuasion. Work on this question has shown
substantial variability in effects, however (see Grush, 1976), and the most common pattern of findings has been
an initial increase in agreement with increased repetition, followed by a decrease in agreement with further
repetition (e.g., Cacioppo & Petty, 1979b). As noted earlier, these effects can be accounted for by a two-stage
model of increases in objective processing of message content, followed by increases in biased processing as
tedium and irritation sets in with further repetitions (see Stewart & Pechman, 1990, for a review).
That is, at low levels of repetition, increasing the number of times a person receives the message provides
the person with greater opportunity to scrutinize the merits of the object or position. This would lead to
increased persuasion if the arguments are strong (as in most of the research noted above), but would actually lead
to decreased persuasion if the arguments are weak (Cacioppo & Petty, 1989). This increase in processing with
increased repetitions should be most evident when the messages are relatively complex (and the number of
repetitions necessary for tedium might be related to the amount of complexity in the messages).
Thought-listing studies have supported the biased-processing interpretation of the tedium effect obtained
with high levels of message repetition. For example, Cacioppo and Petty (1979b) found that counterarguing of
strong arguments decreased from one to three message repetitions (consistent with increases in relatively
objective processing with moderate repetition), but increased from three to five repetitions of the same message
(consistent with irritation or reactance setting in with high levels of repetition). Favorable thoughts showed the
opposite quadratic pattern, and thoughts such as complaints of boredom increased with higher message repetition.
Although increases in repetition led to increases in recall of message arguments (suggesting increased
processing), it was favorability of thoughts rather than amount of recall that predicted agreement. Similar effects
have been found in field settings. For example, Gorn and Goldberg (1980) found that children preferred an ice
cream product most with moderate repetitions of an ad for the ice cream, and that expressions of displeasure
(e.g., ?not again!”) increased with high levels of ad repetition.
Summary of Effects of Persuasion Variables
Although it is still the case that most studies have examined just one role for any given persuasion
variable, each of the possible multiple roles for persuasion variables has been observed for one or more of the
source, message, recipient, and context variables that were reviewed. For example, effects of source
characteristics have been found across the entire elaboration continuum. Even though the most studied role of
source factors is that of persuasion cue when the elaboration likelihood is low, source characteristics have also
been shown to influence persuasion when elaboration likelihood is high, but this has occurred by biasing the
processing of the communication -- especially when arguments in the message are ambiguous. Finally, numerous
source characteristics have been shown to influence the amount of processing of persuasive messages when
background factors do not constrain elaboration to be extremely high or low. Similar multiple roles have been
observed for message, recipient, and context factors. Of course, some variables have only been shown to operate
in one role in existing research, and studies that examine all roles within the same study are quite rare. Future
research will undoubtedly more fully explore the multiple roles by which variables can influence attitudes. In
addition, use of the multiple roles framework might be used to generate unique hypotheses. For example,
although distraction has been studied mostly in its role as a disrupter of processing, it is possible that mild
distraction might also serve as a motivator of processing if people feel that because of the distraction they need to
exert extra effort in attending to the message. If this effort is greater than is actually necessary to overcome the
distraction, enhanced processing over no distraction conditions would be the result. In addition, some forms of
distraction could also be annoying, which could serve as a simple disagreement cue or serve to bias processing
(much as annoying message repetition has been shown to bias processing).
CONSEQUENCES OF DIFFERENT PERSUASION PROCESSES
Understanding the processes by which source, message, recipient, and context variables have their impact
on attitude change is important not only for the conceptual understanding it provides, but also because there are
48
some very important consequences associated with the process by which attitudes are changed. In this final
section, some of the important characteristics and consequences of attitudes that are changed by high- versus loweffort processes are reviewed briefly (see also Eagly & Chaiken, this volume). In particular, the temporal
persistence of attitude changes, the resistance of newly changed attitudes to counterpersuasion, and the ability of
newly formed or changed attitudes to predict behavior (and behavioral intentions) are highlighted.
Over the past few decades, a number of studies have addressed these topics. A general conclusion is that
attitude changes that are accompanied by high levels of issue-relevant cognitive activity about the dimensions
central to the attitude object are stronger than changes that are accompanied by little issue-relevant thought, or
considerable thought but along dimensions that are not central to the merits of the attitude object (see Petty,
Haugtvedt, & Smith, 1995, for a review). High levels of issue-relevant cognitive activity are likely to require
frequent accessing of the attitude and the corresponding knowledge structure. This activity should therefore tend
to increase the number of linkages and strengthen the associations among the structural elements, making the
attitude schema more internally consistent, accessible, and enduring (Crocker, Fiske, & Taylor, 1984; Fazio,
Sanbonmatsu, Powell, & Kardes, 1986; McGuire, 1981). In comparison, attitude change that results from simple
on-line inference or heuristic processes typically involve accessing the attitude structure only once in order to
incorporate the affect or inference associated with a salient persuasion cue (Petty & Cacioppo, 1986a). In general
then, these attitudes should be weaker.
In the ELM analysis, attitude changes are stronger the more they are based on issue-relevant thinking and
it does not matter if this thinking occurs because the experimenter instructs the person to generate a message, if
the thinking is inspired naturally by the personal importance of the issue or an impending discussion, and so
forth. Similarly, it doesn’t matter if the enhanced thinking is relatively objective or is biased by consistency,
reactance, self-esteem, impression management, or other motives. It is important to note, however, that this
consequences postulate does not imply that thoughtfully changed attitudes necessarily will be stronger than one’s
initial attitude. It simply means that a thoughtfully changed attitude will be stronger than an unthoughtfully
changed attitude. When one’s attitude toward some object is changed to a new position, the “old” attitude may
still exist along with the new attitude, and if the new attitude is not as strong as the old one, the old attitude could
still be accessed and guide thinking and behavior. This is particularly important in many applied domains in
which one wants the new attitude rather than the old one to guide action. For example, based on a health
education campaign, adolescents may develop new attitudes toward safe sex. However, even if these new
attitudes are induced by the central route, people may still need to think before they act so that the new attitude
rather the old attitude or salient situational cues guide behavior (Petty, Gleicher, & Jarvis, 1993).
Persistence of Attitude Change
In an attitude change context, persistence refers to the extent to which the newly changed attitude
endures over time. In a comprehensive review of the experimental work on the persistence of attitude change,
Cook and Flay (1978) concluded quite pessimistically that most of the laboratory studies on attitude change
tended to find very little persistence. In the years since this influential paper, it has become more clear when
attitude changes will persist and when they will not.
Current research is compatible with the view that when attitude changes are based on extensive issuerelevant thinking, they tend to endure (e.g., Mackie, 1987; see Petty et al., 1995). That is, conditions that foster
people’s motivation and ability to engage in issue-relevant cognitive activity at the time of message exposure are
associated with increased persistence of persuasion. Thus, research has shown that self-generation of arguments
(e.g., Elms, 1966; Watts, 1967) and autobiographical instances relevant to an issue (Lydon, Zanna, & Ross,
1988), using interesting or involving communication topics (Ronis, Baumgardner, Leippe, Cacioppo, &
Greenwald, 1977), providing increased time to think about a message (e.g., Mitnick & McGinnies, 1958),
increasing message repetition (e.g., Johnson & Watkins, 1971), reducing distraction (e.g., Watts & Holt, 1979)
and leading recipients to believe that they might have to explain or justify their attitudes to other people (e.g.,
Boninger, Brock, Cook, Gruder, & Romer, 1990; Chaiken, 1980) are all associated with increased persistence.
Also, people who characteristically enjoy thinking (high need for cognition) show greater persistence of attitude
change than people who do not (Haugtvedt & Petty, 1992; Verplanken, 1991). Interestingly, simple cues can
become associated with persistent attitudes if the cues remain salient over time (though attitudes that persist for
this reason might often be relatively easy to change when challenged, see Haugtvedt, Schumann, Schneier, &
49
Warren, 1994). Relative persistence can be accomplished by repeated pairings of the cue and attitude object so
that the cue remains relatively accessible (e.g., Haugtvedt et al., 1994) or by reintroducing the cue at the time of
attitude assessment (e.g., Kelman & Hovland, 1953).
The Yale group explicitly acknowledged the role of "peripheral cues" and their impact on attitude
persistence in their work on the "sleeper effect." A sleeper effect is said to occur when a message that is
accompanied initially by a negative cue (e.g., a noncredible source) increases in effectiveness over time (see
Cook, Gruder, Hennigan, & Flay, 1979; Hovland et al., 1949). To account for this effect, Kelman and Hovland
(1953) proposed that in addition to message arguments, various cues could have an impact on attitude change.
These cues were thought to add to (or subtract from) the effects of the persuasive message. Importantly, the cues
and message were viewed as independent and were postulated to have different decay functions. Given this
formulation, a sleeper effect would be produced if a person was exposed to a message with a discounting cue and
the following conditions were met: (a) the message alone had a strong positive impact, (b) the discounting cue
was sufficiently negative to suppress the positive impact of the message, and c) the message conclusion became
dissociated from the discounting cue more quickly than it became dissociated from the message arguments (Cook
et al., 1979). Thus, at a later point in time, it is possible for the positive residue of the message to outlast the
negative effect of the cue, leading to increased agreement with the message conclusion (compared with a nomessage control).
This analysis suggests that one key to producing a sleeper effect is to construct a situation in which both
a strong negative cue and strong arguments have an initial impact. According to the multi-process models,
however, this should be difficult to produce either because of a tradeoff between the impact of central and
peripheral processes (as argument impact increases, cue impact decreases; Petty et al., 1981), or because of an
attenuation effect (processing strong arguments should overpower the negative cue; Maheswaran & Chaiken,
1991), or an interaction between cues and arguments (e.g., people might ignore a message from a low credible
source, overturning one of the critical conditions for the effect; Heesacker et al., 1983). A clever solution to this
conceptual dilemma is to have subjects process the message arguments first so that the strength of the issuerelevant information is realized, and following this present a discounting cue that causes subjects to doubt the
validity of the message (e.g., telling people the message was false). In fact, this is the procedure used in a
number of successful sleeper effect studies (Cook et al., 1979; Kelman & Hovland, 1953). In a relevant series of
experiments, Pratkanis et al. (1988) showed that presenting the discounting cue after the message was critical for
obtaining a reliable sleeper effect (see also Petty et al., 1993).
Resistance to Counterpersuasion
Resistance refers to the extent to which an attitude change is capable of surviving an attack from contrary
information. Attitudes are more resistant the less they change in the direction of contrary information when
challenged. Although attitude persistence and resistance tend to co-occur, their potential independence is shown
conclusively in McGuire's (1964) work on cultural truisms. Truisms such as "you should sleep 8 hours each
night," tend to be highly persistent in a vacuum, but very susceptible to influence when challenged. As McGuire
notes, people have very little practice in defending these beliefs because they have never been attacked. These
beliefs were likely formed with little issue-relevant thinking at a time during childhood when extensive thinking
was relatively unlikely. Instead, the truisms were probably presented repeatedly by powerful, likable, and expert
sources. As noted above, the continual pairing of a belief with positive cues can produce a relatively persistent
attitude, but these attitudes might not prove resistant when attacked.
The resistance of attitudes can be improved by bolstering them with relevant information (e.g., Lewan &
Stotland, 1961). In his work on inoculation theory, McGuire (1964) demonstrated that two kinds of bolstering
can be effective in inducing resistance. One form relies on providing individuals with a supportive defense of
their attitudes or having them generate supportive information. For example, subjects whose initial attitudes
were bolstered by recalling autobiographical instances relevant to the attitude showed greater resistance to an
attacking message than subjects whose attitudes were followed by the generation of autobiographical instances
that were irrelevant to the attitude issue (Ross, McFarland, Conway, & Zanna, 1983). A second type of defense
relies on a biological analogy. That is, McGuire suggested that just as people can be made more resistant to a
disease by giving them a mild form of the germ, people could be made more resistant to discrepant messages by
inoculating their initial attitudes. The inoculation treatment consists of exposing people to a few pieces of
50
counterattitudinal information prior to the threatening communication and showing them how to refute this
information. This presumably produces subsequent resistance because the inoculation poses a threat that
motivates and enables people to develop bolstering arguments for their somewhat weakened attitude (see also
McGuire & Papageorgis, 1961; Pfau et al., 1990).
There is relatively little work on the specific qualities that render attitude changes resistant to attack.
However, any treatment that links the new attitude to the various factors known to be associated with strength
(e.g., confidence, high knowledge, etc.; Petty & Krosnick, 1995; see Eagly & Chaiken, this volume) should
increase the resistance of the attitudes. The existing data support the view that attitudes are more resistant to
attack when they are accessible (Bassili & Fletcher, 1991) and have resulted from considerable issue-relevant
elaboration. For example, Haugtvedt and Petty (1992) provided subjects who were high or low in need for
cognition with an initial message about the safety of a food additive. This initial message, containing strong
arguments from an expert source, was followed by an opposite message containing rather weak arguments from a
different expert source. Although both high and low need for cognition individuals were equally persuaded by
the initial message, the attitudes of the high need for cognition subjects were more resistant to the attacking
message. In addition, high need for cognition individuals engaged in greater counterarguing of the attacking
message. Similarly, as noted previously, Haugtvedt and Wegener (1994) found that subjects who encountered an
initial message under conditions of high personal relevance (and thus processed it extensively) were relatively
uninfluenced by a subsequent opposing message. In comparison, subjects who received the same messages under
conditions of low relevance were more influenced by the second communication. Furthermore, subjects in the
high relevance conditions engaged in more counterarguing of the second (opposing) message.
A strong initial attitude should prove especially effective in resisting a subsequent message if that
message is susceptible to counterarguing (i.e., presents a counterattitudinal position with weak or mixed
arguments). Consistent with this reasoning, Wu and Shaffer (1987) varied attitude strength by manipulating
whether the initial attitude toward a new consumer product was based on direct or indirect experience. They
found that attitudes based on direct experience were more resistant to a counterattitudinal appeal but more
susceptible to a proattitudinal appeal than were attitudes based on indirect experience. Information gleaned from
direct experience (e.g., the taste of the product) might be more accessible, held with greater confidence, and be
linked more strongly to the attitude object allowing a person to more easily recognize the flaws in contrary
information but the merits in congruent information. In addition, when initial attitudes were based on direct
experience, attitudes were less influenced by the credibility of the source of the second message regardless of
whether the message was pro or counterattitudinal. That is, strong attitudes were less susceptible to peripheral
cues (see Petty et al., 1995, for a review).
Attitude-Behavior Consistency
Perhaps the most important quality of attitudes for those interested in applications of persuasion theory
concerns the ability of attitudes to predict peoples' actions. A number of situational and dispositional factors
have been shown to enhance the consistency of attitudes with behaviors (see Kraus, 1995, for a recent metaanalysis). For example, attitudes are more predictive of behavior when: (a) the persons tested are of a certain
personality type (e.g., are low in "self-monitoring," Snyder & Swann, 1976; or high in "need for cognition,"
Cacioppo, Petty, Kao, & Rodriguez, 1986), (b) the attitudes in question are consistent with underlying beliefs
(e.g., Norman, 1975); c) the attitudes are based on high rather than low amounts of issue-relevant knowledge
(e.g., Davidson, Yantis, Norwood, & Montano, 1985; Kallgren & Wood, 1986); (d) the attitudinal issues are on
topics of high personal relevance (e.g., Petty et al., 1983; Verplanken, 1991); and (e) the attitudes are high rather
than low in accessibility (e.g., Bassili, 1993, 1995; Fazio & Herr, 1983). What these factors have in common is
that each is associated with attitudes that are likely based on high amounts of issue-relevant thinking (see Petty et
al., 1995, for a review).
Interestingly, if people are asked to think about the basis of their attitudes just prior to attitude
measurement, attitude-behavior consistency could be reduced if thinking produces an expressed attitude that is
not representative of the true one (Wilson et al, 1989). For example, if the central merits of an attitude object are
affectively based, but a thinking task makes cognitive rather than affective information salient prior to attitude
expression, the attitude expressed after thought will be less predictive of behavior than an attitude expressed
without thought -- especially if the behavior is affectively based as well (see Millar & Tesser, 1992, for a review).
51
Although much research has examined how methodological factors and existing characteristics of
attitudes (e.g., extent of knowledge), people (e.g., personality), and situations (e.g., time pressure) moderate
attitude-behavior consistency, relatively few studies have examined whether different attitude formation or
change processes are related to the ability of newly formed or changed attitudes to predict behavior. However,
some research has shown that attitudes correlate to a greater extent with behavior when the attitudes were formed
under high than under low personal relevance conditions (Leippe & Elkin, 1987; Sivacek & Crano, 1982). Fazio
and his colleagues have examined attitudes that were formed as a result of direct or indirect experience with the
object and have found that the former are more predictive of behavior (see Fazio & Zanna, 1981, for a review).
A primary reason for this is that attitudes based on direct experience are more accessible and thus more able to
color perception of the attitude object and guide behavior (Fazio, 1990, 1995; see Eagly & Chaiken, this volume,
for additional discussion). One possible reason why attitudes based on direct experience are more accessible is
that direct experience might typically lead to greater thought relevant to the attitude object than passive exposure
to a persuasive message.
CONCLUSIONS
Research on attitude change has come a long way from initial assumptions that variables have
unidirectional effects on persuasion (e.g., source credibility is good for persuasion), or that variables have an
impact on persuasion by a single process (e.g., source credibility facilitates learning of the message). It is now
clear that the many source, message, recipient, and context variables that have been studied over the past century
can have complex effects -- increasing persuasion in some situations and decreasing it in others. These
bidirectional effects have been the case even for variables that on the surface, at least, seemed to be “obviously”
unidirectional. For example, what could be more obvious but that distraction would be detrimental to persuasion
or that expert sources would be good for persuasion? Yet, contemporary research indicates that distraction can
enhance persuasion if the arguments are weak because the distraction can disrupt the normal counterarguing that
would take place and expertise can be bad for persuasion when it leads to enhanced thinking about weak
arguments.
Just as this single effect assumption had to be abandoned, so too did the single process assumption that
variables tend to have their impact on attitudes by just one mechanism. It is now clear that variables can produce
attitude change by different processes in different situations. Thus, for example, a credible source or a positive
mood can lead to more persuasion by invoking a simple heuristic in low-thought situations, by influencing the
extent of message processing when people are unsure whether thought is merited or not, and by biasing the
ongoing processing when thought is high. Furthermore, it is clear that the same outcome can be produced by
different processes in different situations. Thus, a positive mood can produce more persuasion under low
elaboration conditions by a simple inference process, but can produce the same amount of persuasion under high
elaboration conditions by biasing the content of one’s elaborations of the message content.
Although there are multiple specific processes that can determine the extent and direction of attitude
change, current research strongly indicates that it is useful to divide the theoretical processes responsible for
modifying attitudes into those that emphasize effortful thinking about the central merits of the attitude object
from those that do not. This framework allows understanding and prediction of what variables will affect
attitudes by what processes in what general situations and what the consequences of these attitude are. This
framework also helps to place the various mini-theories of attitude change in their proper domain of operation.
For example, a high effort process like cognitive responses should account for attitude change in those contexts
in which thinking is expected to be high, whereas a lower effort process such as balance or use of simple
heuristics should be more likely to account for empirical effects in those contexts in which thinking is expected
to be low. Finally, recognition of an elaboration continuum permits understanding and prediction of the strength
of attitudes changed by different processes. That is, attitudes that are changed as a result of considerable mental
effort tend to be stronger than those changed with little thought and thus are more persistent, resistant to
counterpersuasion, and predictive of behavior than attitudes that are changed by processes invoking little mental
effort in assessing the central merits of the object.
As might be expected, contemporary persuasion theories and those that will dominate in the 21st century
are considerably more complicated than those that reigned in earlier periods. These theories must accommodate
multiple effects of individual variables and multiple processes by which these variables have their impact.
52
Furthermore, these theories must specify the conditions under which the different processes operate, and any
differential consequences of these processes. Attitude change researchers have made great strides over the past
century in identifying the building blocks of such a multi-faceted theory, though considerable work remains to be
done.
REFERENCES
Abelson, R. P. (1995). Attitude extremity. In R. E. Petty & J. A. Krosnick (Eds.). Attitude strength:
Antecedents and consequences (pp. 25-41). Mahwah, NJ: Erlbaum.
Abelson, R. P., Aronson, E., McGuire, W. J., Newcomb, T. N., Rosenberg, M. J., & Tannenbaum, P.
(1968). Theories of cognitive consistency: A sourcebook. Chicago, IL: Rand McNally.
Abelson, R. P., & Lesser, G. S. (1959). The measurement of persuasibility in children. In C. I. Hovland
& I. L. Janis (Eds.), Personality and persuasibility (pp. 141-166). New Haven, CT: Yale University Press.
Ahlering, R. F., & Parker, L. D. (1989). Need for cognition as a moderator of the primacy effect. Journal
of Research in Personality, 23, 313-317.
Allen, M. (1991). Meta-analysis comparing the persuasiveness of one-sided and two-sided messages.
Western Journal of Speech Communication, 55, 390-404.
Allen, M., Hale, J., Mongeau, P., Berkowits-Stafford, S., Stafford, S., Shanahan, W., Agee, P., Dillon, K.,
Jackson, R., & Ray, C. (1990). Testing a model of message sidedness: Three replications. Communication
Monographs, 57, 274-291.
Allport, F. H. (1924). Social psychology. Boston: Houghton Mifflin.
Allport, G. W. (1935). Attitudes. In C. Murchison (Ed.), Handbook of social psychology (pp. 798-884).
Worcester, MA: Clark University Press.
Alwin, D. F., Cohen, R. L., & Newcomb, T. M. (1991). The women of Bennington: A study of political
orientations over the life span. Madison: University of Wisconsin Press.
Anderson, C. A., Lepper, M. R., & Ross, L. (1980). Perseverance of social theories: The role of
explanation in the persistence of discredited information. Journal of Personality and Social Psychology, 39,
1037-1049.
Anderson, N. H. (1971). Integration theory and attitude change. Psychological Review, 78, 171-206.
Anderson, N. (1981). Integration theory applied to cognitive responses and attitudes. In R. Petty, T.
Ostrom, & T. Brock (Ed.), Cognitive responses in persuasion (pp. 361-397). Hillsdale, NJ: Erlbaum.
Anderson, N. H., & Hovland, C. I. (1957). The representation of order effects in communication
research (Appendix A). In C. I. Hovland (Ed.), The order of presentation in persuasion (pp. 158-169). New
Haven, CT: Yale University Press.
Andreoli, V., & Worchel, S. (1978). Effects of media, communicator, and message position on attitude
change. Public Opinion Quarterly, 42, 59-70.
Apsler, R., & Sears, D. O. (1968). Warning, personal involvement, and attitude change. Journal of
Personality and Social Psychology, 9, 162-166.
Areni, C. S., & Lutz, R. J. (1988). The role of argument quality in the elaboration likelihood model.
Advances in Consumer Research, 15, 197-203.
Arkes, H. R., Boehm, L. E., & Xu, G. (1991). Determinants of judged validity. Journal of Experimental
Social Psychology, 27, 576-605.
Aronson, E. (1969). The theory of cognitive dissonance: A current perspective. In L. Berkowitz (Ed.),
Advances in experimental social psychology (Vol. 4, pp. 1-34). San Diego, CA: Academic Press.
Aronson, E., & Mills, J. (1959). The effects of severity of initiation on liking for a group. Journal of
Abnormal and Social Psychology, 59, 177-181.
Aronson, E., Turner, J., & Carlsmith, M. (1963). Communicator credibility and communicator
discrepancy as determinants of opinion change. Journal of Abnormal and Social Psychology, 67, 31-36.
Asch, S. E. (1951). Effects of group pressure upon the modification and distortion of judgments. In H.
Guetzkow (Ed.), Groups, leadership, and men (pp. 177-190). Pittsburgh, PA: Carnegie Press.
Asch, S. E. (1956). Studies of independence and conformity: I. A minority of one against a unanimous
majority. Psychological Monographs, 70 (9, Whole No. 416)
53
Axsom, D., Yates, S. M., & Chaiken, S. (1987). Audience response as a heuristic cue in persuasion.
Journal of Personality and Social Psychology, 53, 30-40.
Bagozzi, R. P. (1985). Expectancy-value attitude models: An analysis of critical theoretical issues.
International Journal of Research in Marketing, 2, 43-60.
Baker, S. M., & Petty, R. E. (1994). Majority and minority influence: Source-position imbalance as a
determinant of message scrutiny. Journal of Personality and Social Psychology, 67, 5-19.
Bargh, J. A., Chaiken, S., Raymond, P., & Hymes, C. (1996). The automatic evaluation effect:
Unconditional automatic attitude activation with a pronunciation task. Journal of Experimental Social
Psychology, 31, 104-128.
Baron, R. S., Baron, P. H., & Miller, N. (1973). The relation between distraction and persuasion.
Psychological Bulletin, 80, 310-323.
Baron, R. S., Inman, M., Kao, C., & Logan, H. (1992). Emotion and superficial social processing.
Motivation and Emotion, 16, 323-345.
Baron, R. S., Logan, H., Lilly, J., Inman, M., & Brennan, M. (1994). Negative emotion and message
processing. Journal of Experimental Social Psychology, 30, 181-201.
Bassili, J. N. (1993). Response latency versus certainty as indexes of the strength of voting intentions in a
cati survey. Public Opinion Quarterly, 57, 54-61.
Bassili, J. N. (1995). Response latency and the accessibility of voting intentions: what contributes to
accessibility and how it affects vote choice. Personality and Social Psychology Bulletin, 21, 686-695.
Bassili, J. N., & Fletcher, J. F. (1991). Response-time measurement in survey research: A method for
CATI and a new look at nonattitudes. Public Opinion Quarterly, 55, 331-346.
Bates, E. (1976). Language and context: The acquisition of pragmatics. New York, NY: Academic Press.
Beauvois, J. L., Bungert, M., & Mariette, P. (1995). Forced compliance: Commitment to compliance and
commitment to activity. European Journal of Social Psychology, 25, 17-26.
Beck, K. H., & Frankel, A. (1981). A conceptualization of threat communications and protective health
behavior. Social Psychology Quarterly, 44, 204-217.
Bem, D. J. (1965). An experimental analysis of self-persuasion. Journal of Experimental Social
Psychology, 1, 199-218.
Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in experimental social
psychology (Vol. 6, pp. 1-62). New York: Academic Press.
Berkowitz, L., & Lundy, R. M. (1957). Personality characteristics related to susceptibility to influence
by peers or authority figures. Journal of Personality, 25, 306-316.
Biek, M., Wood, W., & Chaiken, S. (1996). Working knowledge and cognitive processing: On the
determinants of bias. Personality and Social Psychology Bulletin, 22, 547-556.
Birnbaum, M. H., Wong, R., & Wong, L. K. (1976). Combining information from sources that vary in
credibility. Memory and Cognition, 4, 330-336.
Blaney, P. H. (1986). Affect and memory: A review. Psychological Bulletin, 99, 229-246.
Bless, H., Bohner, G., Schwarz, N., & Strack, F. (1990). Mood and persuasion: A cognitive response
analysis. Personality and Social Psychology Bulletin, 16, 331-345.
Boninger, D. S., Brock, T. C., Cook, T. D., Gruder, C. L., & Romer, D. (1990). Discovery of reliable
attitude change persistence resulting from a transmitter tuning set. Psychological Science, 1, 268-171.
Boninger, D., Krosnick, J. A., Berent, M. K., & Fabrigar, L. R. (1995). The causes and consequences of
attitude importance. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and consequences
(pp. 159-189). Mahwah, NJ: Lawrence Erlbaum Associates.
Bornstein, R. F. (1989). Exposure and affect: Overview and meta-analysis of research, 1968-1987.
Psychological Bulletin, 106, 265-289.
Bornstein, R. F., & D'Agostino, P. R. (1992). Stimulus recognition and the mere exposure effect.
Journal of Personality and Social Psychology, 63, 545-552.
Bornstein, R. F., & D’Agostino, P. R. (1994). The attribution and discounting of perceptual fluency:
Preliminary tests of a perceptual fluency/attributional model of the mere exposure effect. Social Cognition, 12,
103-128.
54
Boster, F. J., & Mongeau, P. (1984). Fear-arousing persuasive messages. In R. N. Bostrom (Ed.),
Communication yearbook (Vol. 8, pp. 330-375). Beverly Hills, CA: Sage.
Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129-148.
Breckler, S. J. (1984). Empirical validation of affect, behavior, and cognition as distinct components of
attitude. Journal of Personality and Social Psychology, 47, 1191-1205.
Breckler, S. J., & Wiggins, E. C. (1991). Cognitive responses in persuasion: Affective and evaluative
determinants. Journal of Experimental Social Psychology, 27, 180-200.
Brehm, J. W. (1956). Postdecision changes in the desirability of alternatives. Journal of Abnormal and
Social Psychology, 52, 384-389.
Brehm, J. W. (1966). A theory of psychological reactance. San Diego, CA: Academic Press.
Brehm, J. W., & Cohen, A. R. (1962). Explorations in cognitive dissonance. New York, NY: Wiley.
Brock, T. C. (1965). Communicator-recipient similarity and decision change. Journal of Personality and
Social Psychology, 1, 650-654.
Brock, T. C. (1967). Communication discrepancy and intent to persuade as determinants of
counterargument production. Journal of Experimental Social Psychology, 3, 296-309.
Brock, T. C. (1968). Implications of commodity theory for value change. In A. G. Greenwald, T. C.
Brock, & T. M. Ostrom (Ed.), Psychological foundations of attitudes (pp. 243-296). New York: Academic Press.
Brucks, M., Armstrong, G. M., & Goldberg, M. E. (1988). Children's use of cognitive defenses against
television advertising: A cognitive response approach. Journal of Consumer Research, 14, 471-482.
Burnkrant, R. E., & Howard, D. J. (1984). Effects of the use of introductory rhetorical questions versus
statements on information processing. Journal of Personality and Social Psychology, 47, 1218-1230.
Burnkrant, R. E., & Unnava, R. (1989). Self-referencing: A strategy for increasing processing of
message content. Personality and Social Psychology Bulletin, 15, 628-638.
Burnstein, E., & Vinokur, A. (1975). What a person thinks upon learning he has chosen differently from
others: Nice evidence for the persuasive-arguments explanation of choice shifts. Journal of Experimental Social
Psychology, 11, 412-426.
Byrne, D., & Griffitt, W. (1966). A developmental investigation of the law of attraction. Journal of
Personality and Social Psychology, 4, 699-702.
Cacioppo, J. T., & Berntson, G. G. (1994). Relationship between attitudes and evaluative space: A
critical review with emphasis on the separability of positive and negative substrates. Psychological Bulletin, 115,
401-423.
Cacioppo, J. T., Gardner, W. L., & Berntsen, G. G. (1997). Beyond bipolar conceptualizations and
measures: The case of attitudes and evaluative space. Personality and Social Psychology Review.
Cacioppo, J. T., Marshall-Goodell, B. S., Tassinary, L. G., & Petty, R. E. (1992). Rudimentary
determinants of attitudes: Classical conditioning is more effective when prior knowledge about the attitude
stimulus is low than high. Journal of Experimental Social Psychology, 28, 207-233.
Cacioppo, J. T., & Petty, R. E. (1979a). Attitudes and cognitive response: An electro-physiological
approach. Journal of Personality and Social Psychology, 37, 2181-2199.
Cacioppo, J. T., & Petty, R. E. (1979b). Effects of message repetition and position on cognitive
responses, recall, and persuasion. Journal of Personality and Social Psychology, 37, 97-109.
Cacioppo, J. T., & Petty, R. E. (1980). Sex differences in influenceability: Toward specifying the
underlying processes. Personality and Social Psychology Bulletin, 6, 651-656.
Cacioppo, J. T., & Petty, R. E. (1981). Social psychological procedures for cognitive response
assessment: The thought listing technique. In T. Merluzzi, C. Glass, & M. Genest (Ed.), Cognitive assessment
(pp. 309-342). New York: Guilford.
Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality and Social
Psychology, 42, 116-131.
Cacioppo, J. T., & Petty, R. E. (1989). Effects of message repetition on argument processing, recall, and
persuasion. Basic and Applied Social Psychology, 10, 3-12.
Cacioppo, J.T., Petty, R.E., Feinstein, J., & Jarvis, B. (1996). Individual differences in cognitive
motivation: The life and times of people varying in need for cognition. Psychological Bulletin, 119, 197-253.
55
Cacioppo, J. T., Petty, R. E., Kao, C. F., & Rodriquez, R. (1986). Central and peripheral routes to
persuasion: An individual difference perspective. Journal of Personality and Social Psychology, 51, 1032-1043.
Cacioppo, J. T., Petty, R. E., & Morris, K. J. (1983). Effects of need for cognition on message
evaluation, recall, and persuasion. Journal of Personality and Social Psychology, 45, 805-818.
Cacioppo, J. T., Petty, R. E., & Sidera, J. (1982). The effects of salient self-schema on the evaluation of
proattitudinal editorials: Top-down versus bottom-up message processing. Journal of Experimental Social
Psychology, 18, 324-338.
Cacioppo, J. T., Priester, J. R., & Berntson, G. G. (1993). Rudimentary determinants of attitudes II: Arm
flexion and extension have differential effects on attitudes. Journal of Personality and Social Psychology, 65, 517.
Calder, B. J., Ross, M., & Insko, C. A. (1973). Attitude change and attitude attribution: Effects of
incentive, choiice, and consequences. Journal of Personality and Social Psychology, 25, 84-99.
Carver, C. S., & Scheier, M. F. (1982). Outcome expectancy, locus of attribution for expectancy, and
self-directed attention as determinants of evaluations and performance. Journal of Experimental Social
Psychology, 18, 184-200.
Ceci, S. J., & Bruck, M. (1993). Suggestibility of the child witness: A historical review and synthesis.
Psychological Bulletin, 113, 403-439.
Chaiken, S. (1978). Unpublished doctoral dissertation. University of Massachusetts, Amherst, MA.
Chaiken, S. (1980). Heuristic versus systematic information processing in the use of source versus
message cues in persuasion. Journal of Personality and Social Psychology, 39, 752-766.
Chaiken, S. (1986). Physical appearance and social influence. In C. P. Herman, M. P. Zanna, & E. T.
Higgins (Eds.), Physical appearance, stigma and social behavior: The Ontario symposium (Vol. 3). Hillsdale, NJ:
Erlbaum.
Chaiken, S. (1987). The heuristic model of persuasion. In M. P. Zanna, J. M. Olson, & C. P. Herman
(Ed.), Social influence: The Ontario symposium (Vol. 5, pp. 3-39). Hillsdale, NJ: Erlbaum.
Chaiken, S., & Baldwin, M. W. (1981). Affective-cognitive consistency and the effect of salient
behavioral information on the self-perception of attitudes. Journal of Personality and Social Psychology, 41,
1-12.
Chaiken, S., & Eagly, A. H. (1976). Communication modality as a determinant of message
persuasiveness and message comprehensibility. Journal of Personality and Social Psychology, 34, 605-614.
Chaiken, S., & Eagly, A. H. (1983). Communication modality as a determinant of persuasion: The role
of communicator salience. Journal of Personality and Social Psychology, 45, 241-256.
Chaiken, S., Pomerantz, E. M., & Giner-Sorolla, R. (1995). Structural consistency and attitude strength.
In R. E. Petty and J. A. Krosnick (Eds.) Attitude strength: Antecedants and consequences (pp. 387-412).
Mahwah, NJ: Erlbaum.
Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic processing within and
beyond the persuasion context. In J. S. Uleman, & J. A. Bargh (Ed.), Unintended thought (pp. 212-252). New
York: Guilford Press.
Chaiken, S., & Maheswaran, D. (1994). Heuristic processing can bias systematic processing: Effects of
source credibility, argument ambiguity, and task importance on attitude judgment. Journal of Personality and
Social Psychology, 66, 460-473.
Chaiken, S., Wood, W., & Eagly, A. H. (1996). Principles of persuasion. In E. T. Higgins, A. W.
Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 702-742). New York: Guilford Press.
Chaiken, S., & Yates, S. M. (1985). Affective-cognitive consistency and thought-induced attitude
polarization. Journal of Personality and Social Psychology, 49, 1470-1481.
Chattopadhyay, A., & Alba, J. W. (1988). The situational importance of recall and inference in
consumer decision making. Journal of Consumer Research, 15, 1-12.
Chen, H. C., Reardon, R., Rea, C., & Moore, D. J. (1992). Forewarning of content and involvement:
Consequences for persuasion and resistance to persuasion. Journal of Experimental Social Psychology, 28, 523541.
Cialdini, R. B. (1987). Compliance principles of compliance professionals: Psychologists of necessity.
56
In M. P. Zanna, J. M. Olson, & C. P. Herman (Ed.), Social influence: The Ontario symposium (Vol. 5, pp.
165-184). Hillsdale, NJ: Erlbaum.
Cialdini, R. B. (this volume). Social influence, conformity, and compliance. In D. Gilbert, S. Fiske, &
G. Lindzey (Eds.) Handbook of social psychology.
Cialdini, R. B., & Petty, R. E. (1981). Anticipatory opinion effects. In R. E. Petty, T. M. Ostrom, & T. C.
Brock (Eds.), Cognitive responses in persuasion (pp. 217-235). Hillsdale, NJ: Erlbaum.
Cialdini, R.B., Trost, M.R., & Newsom, J. T. (1995). Preference for consistency: The development of a
valid measure and the discovery of surprising behavioral implications. Journal of Personality and Social
Psychology, 69, 318-328.
Cohen, A. R. (1957). Need for cognition and order of communication as a determinant of opinion
change. In C. I. Hovland (Ed.), Order of presentation in persuasion (pp. 79-97). New Haven, CT: Yale University
Press.
Converse, P. E. (1970). Attitudes and non-attitudes: Continuation of a dialogue. In E. R. Tufte (Ed.),
The quantitative analysis of social problems (pp. 168-189). Reading, MA: Addison-Wesley.
Cook, T. D., & Flay, B. R. (1978). The persistence of experimentally induced attitude change. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 11, pp. 1-57). New York: Academic Press.
Cook, T. D., Gruder, C. L., Hennigan, K. M., & Flay, B. R. (1979). History of the sleeper effect: Some
logical pitfalls in accepting the null hypothesis. Psychological Bulletin, 86, 662-679.
Cooper, E., & Dinerman, H. (1951). Analysis of the film ?Don’t Be a Sucker”: A study of
communication. Public Opinion Quarterly, 15, 243-264.
Cooper, H. M. (1979). Statistically combining independent studies: Meta-analysis of sex differences in
conformity research. Journal of Personality and Social Psychology, 37, 131-146.
Cooper, J., & Fazio, R. H. (1984). A new look at dissonance theory. In L. Berkowitz (Ed.), Advances in
experimental social psychology (Vol. 17, pp. 229-266). New York: Academic Press.
Cooper, W. H. (1981). Ubiquitous halo. Psychological Bulletin, 90, 218-224.
Cox, D. S., & Cox, A. D. (1988). What does familiarity breed: Complexity as a moderator of repetition
effects in advertisement evaluation. Journal of Consumer Research, 15, 111-116.
Crano, W. D. (1995). Attitude strength and vested interest. In R. E. Petty & J. A. Krosnick (Eds.).
Attitude strength: Antecedents and consequences (pp. 131-157). Mahwah, NJ: Erlbaum.
Crano, W. D., & Hannula-Bral, K. A. (1994). Context/categorization model of social influence: Minority
and majority influence in the formation of a novel response norm. Journal of Experimental Social Psychology,
30, 247-276.
Crites, S. L., Jr., Fabrigar, L. R., & Petty, R. E. (1994). Measuring the affective and cognitive properties
of attitudes: Conceptual and methodological issues. Personality and Social Psychology Bulletin, 20, 619-634.
Crocker, J., Fiske, S. T., & Taylor, S. E. (1984). Schematic bases of belief change. In R. Eiser (Ed.),
Attitudinal judgment. New York: Springer-Verlag.
Cruchfield, R. S. (1955). Conformity and character. American Psychologist, 10, 191-198.
Davidson, A. R., Yantis, S., Norwood, M., & Montano, D. E. (1985). Amount of information about the
attitude object and attitude-behavior consistency. Journal of Personality and Social Psychology, 49, 1184-1198.
Deaux, K., & LaFrance, M. (this volume). Gender. In D. Gilbert, S. Fiske, & G. Lindzey (Eds.)
Handbook of social psychology.
DeBono, K. G. (1987). Investigating the social-adjustive and value-expressive functions of attitudes:
Implications for persuasion processes. Journal of Personality and Social Psychology, 52, 279-287.
DeBono, K. G., & Harnish, R. J. (1988). Source expertise, source attractiveness, and the processing of
persuasive information: A functional approach. Journal of Personality and Social Psychology, 55, 541-546.
DeBono, K. G., & Klein, C. (1993). Source expertise and persuasion: The moderating role of recipient
dogmatism. Personality and Social Psychology Bulletin, 19, 167-173.
De Houwer, J., Baeyens, F., & Eelen, P. (1994). Verbal evaluative conditioning with undetected US
presentations. Behavior Research and Therapy,. 32, 629-633.
Deci, E. L. (1975). Intrinsic motivation. New York: Plenum.
Delia, J. G. (1987). Communication research: A history. In C. R. Berger & S. H. Chaffee (Eds.),
57
Handbook of communication science (pp. 20-98). Newbury Park, CA: Sage.
Dillard, J. P. (in press). Rethinking the study of fear appeals. Communication Theory.
Dillehay, R. C., Insko, C. A., & Smith, M. B. (1966). Logical consistency and attitude change. Journal of
Personality and Social Psychology, 3, 646-654.
Ditto, P. H., Jemmott, J. B., & Darley, J. M. (1988). Appraising the threat of illness: A mental
representational approach. Health Psychology, 7, 183-200.
Ditto, P. H., & Lopez, D. F. (1992). Motivated skepticism: Use of differential decision criteria for
preferred and nonpreferred conclusions. Journal of Personality and Social Psychology, 63, 568-584.
Doob, L. W. (1935). Propaganda. New York: Holt, Rinehart & Winston.
Downing, J. W., Judd, C. M., & Brauer, M. (1992). Effects of repeated expressions on attitude
extremity. Journal of Personality and Social Psychology, 63, 17-29.
Eagly, A. H. (1974). Comprehensibility of persuasive arguments as a determinant of opinion change.
Journal of Personality and Social Psychology, 29, 758-773.
Eagly, A. H. (1978). Sex differences in influenceability. Psychological Bulletin, 85, 86-116.
Eagly, A. H. (1981). Recipient characteristics as determinants of responses to persuasion. In R. E. Petty,
T. M. Ostrom, & T. C. Brock (Eds.), Cognitive responses in persuasion (pp. 173-195). Hillsdale, NJ: Erlbaum.
Eagly, A. H., & Chaiken, S. (1984). Cognitive theories of persuasion. In L. Berkowitz (Ed.), Advances
in experimental social psychology (Vol. 17, pp. 268-361. New York: Academic Press..
Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Forth Worth, TX: Harcourt, Brace,
Jovanovich.
Eagly, A. H., & Chaiken, S. (this volume). Attitude structure and function. In D. Gilbert, S. Fiske, & G.
Lindzey (Eds.), The handbook of social psychology (4th edition).
Eagly, A. H., Chaiken, S., & Wood, W. (1981). An attribution analysis of persuasion. In J. H. Harvey,
W. J. Ickes, & R. F. Kidd (Ed.), New direction in attribution research (Vol. 3, pp. 37-62). Hillsdale, NJ: Erlbaum.
Eagly, A. H., & Telaak, K. (1972). Width of the latitude of acceptance as a determinant of attitude
change. Journal of Personality and Social Psychology, 23, 388-397.
Eagly, A. H., & Warren, R. (1976). Intelligence, comprehension, and opinion change. Journal of
Personality, 44, 226-242.
Eagly, A. H., & Wood, W. (1991). Explaining sex differences in social behavior: A meta-analytic
perspective. Personality and Social Psychology Bulletin, 17, 306-315.
Eagly, A. H., Wood, W., & Chaiken, S. (1978). Causal inferences about communicators and their effect
on opinion change. Journal of Personality and Social Psychology, 36, 424-435.
Edwards, K. (1990). The interplay of affect and cognition in attitude formation and change. Journal of
Personality and Social Psychology, 59, 202-216.
Edwards, K., & Smith, E. E. (1996). A disconfirmation bias in the evaluation of arguments. Journal of
Personality and Social Psychology, 71, 5-24.
Edwards, K., & von Hippel, W. (1995). Hearts and minds: The priority of affective versus cognitive
factors in person perception. Personality and Social Psychology Bulletin, 21, 996-1011.
Elkin, R. A., & Leippe, M. R. (1986). Physiological arousal, dissonance, and attitude change: Evidence
for a dissonance-arousal link and a "Don't remind me" effect. Journal of Personality and Social Psychology, 51,
55-65.
Elliot, A. J., & Devine, P. G. (1994). On the motivational nature of cognitive dissonance: Dissonance as
psychological discomfort. Journal of Personality and Social Psychology, 67, 382-394.
Elms, A. C. (1966). Influence of fantasy ability on attitude change through role-playing. Journal of
Personality and Social Psychology, 4, 36-43.
Eysenck, H. J., & Furneaux (1945). Primary and secondary suggestibility: an experimental and statistical
study. Journal of Experimental Psychology, 35, 485-503.
Fabrigar, L. R., Wegener, D. T., Priester, J. R., & Petty, R. E. (1995, May). The impact of attitude
accessibility on cognitive elaboration of persuasive messages. Paper presented at the annual meeting of the
Midwestern Psychological Association, Chicago, IL.
Fazio, R. H. (1986). How do attitudes guide behavior? In R. M. Sorrentino, & E. T. Higgins (Ed.),
58
Handbook of motivation and cognition: Foundations of social behavior (pp. 204-243). New York: Guilford
Press.
Fazio, R. H. (1990). Multiple processes by which attitudes guide behavior: The model model as an
integrative framework. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 23, pp. 75-109).
San Diego, CA: Academic Press.
Fazio, R. H. (1995). Attitudes as object-evaluation associations: Determinants, consequences, and
correlates of attitude accessibility. In R. E. Petty, & J. A. Krosnick (Ed.), Attitude strength: Antecedents and
consequences (pp. 247-282). Mahwah, NJ: Erlbaum.
Fazio, R. H., & Cooper, J. (1983). Arousal in the dissonance process. In J. T. Cacioppo, & R. E. Petty
(Ed.), Social psychophysiology: A sourcebook (pp. 122-152). New York: Guilford Press.
Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). Variability in automatic activation
as an unobtrusive measure of racial attitudes: A bona fide pipeline? Journal of Personality and Social
Psychology, 69, 1013-1027.
Fazio, R. H., Sanbonmatsu, D. M., Powell, M. C., & Kardes, F. R. (1986). On the automatic activation of
attitudes. Journal of Personality and Social Psychology, 50, 229-238.
Fazio, R. H., & Williams, C. J. (1986). Attitude accessibility as a moderator of the attitude-perception
and attitude-behavior relations: An investigation of the 1984 presidential election. Journal of Personality and
Social Psychology, 51, 505-514.
Fazio, R. H., & Zanna, M. P. (1981). Direct experience and attitude behavior consistency. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 14, pp. 161-202). New York: Academic
Press.
Fazio, R. H., Zanna, M. P., & Cooper, J. (1977). Dissonance and self-perception: An integrative view of
each theory's proper domain of application. Journal of Experimental Social Psychology, 13, 464-479.
Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row, Peterson.
Festinger, L. (1964). Conflict, decision, and dissonance. Stanford, CA: Stanford University Press.
Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forced compliance. Journal of
Abnormal and Social Psychology, 58, 203-210.
Festinger, L., & Maccoby, N. (1964). On resistance to persuasive communication. Journal of Abnormal
and Social Psychology, 68, 359-366.
Festinger, L., & Thibaut, J. (1951). Interpersonal communication in small groups. Journal of Abnormal
and Social Psychology, 46, 92-100.
Fine, B. J. (1957). Conclusion-drawing, communicator credibility, and anxiety as factors in opinion
change. Journal of Abnormal and Social Psychology, 54, 369-374.
Fink, E. L., Rey, L.D., Johnson, K.W., Spenner, K., Morton, D.R., & Flores, E. T. (1975). The effects of
family occupational type, sex, and appeal style on helping behavior. Journal of Experimental Social Psychology,
11, 43-52.
Fishbein, M., & Ajzen, I. (1972). Attitudes and opinions. Annual Review of Psychology, 23, 487-544.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior. Reading, MA:
Addison-Wesley.
Fishbein, M., & Ajzen, I. (1981). Acceptance, yielding and impact: Cognitive processes in persuasion.
In R. E. Petty, T. M. Ostrom, & T. C. Brock (Ed.), Cognitive responses in persuasion (pp. 339-359). Hillsdale,
NJ: Erlbaum.
Fishbein, M., & Middlestadt, S. (1995). Noncognitive effects on attitude formation and change: Fact or
artifact? Journal of Consumer Psychology, 4, 181-202.
Fiske, S.T. (1980). Attention and weight in person perception: The impact of negative and extreme
behavior. Journal of Personality and Social Psychology, 38, 889-906.
Fiske, S. T., Morling, B., & Stevens, L. E. (1996). Controlling self and others: A theory of anxiety,
mental control, and social control. Personality and Social Psychology Bulletin, 22, 115-123.
Fiske, S. T., & Neuberg, S. L. (1990). A continuum of impression formation, from category-based to
individuating processes: Influences of information and motivation on attention and interpretation. In M. P. Zanna
(Ed.), Advances in experimental social psychology (Vol 23, pp. 1-74). San Diego, CA: Academic Press.
59
Fleming, D. (1967). Attitude: History of a concept. Perspectives in American History, 1, 287-365.
Förster, J., & Strack, F. (1996). Influence of overt head movements on memory for valenced words: A
case of conceptual-motor compatibility. Journal of Personality and Social Psychology, 71, 421-430.
Fogelin (1982). Understanding arguments (2nd ed.). New York: Harcourt, Brace, Jovanovich.
Freedman, J., & Sears, D. (1965). Warning, distraction, and resistance to influence. Journal of
Personality and Social Psychology, 1, 262-266.
Freedman, J. L., Sears, D. O., & O’Connor, E. F. (1964). The effects of anticipated debate and
commitment on the polarization of audience opinion. Public Opinion Quarterly, 28, 615-627.
French, J. R. P., & Raven, B. (1959). The bases of social power. In D. Cartwright (Ed.), Studies in social
power (pp. 150-167). Ann Arbor, MI: University of Michigan.
Fried, C. B., & Aronson, E. (1995). Hypocrisy, misattribution, and dissonance reduction. Personality
and Social Psychology Bulletin, 21, 925-933.
Friedrich, J., Fetherstonhaugh, D., Casey, S., & Gallagher, D. (1996). Argument integration and attitude
change: Suppression effects in the integration of one-sided arguments that vary in persuasiveness. Personality
and Social Psychology Bulletin, 22, 179-191.
Fukada, H. (1986). Psychological processes mediating the persuasion inhibiting effect of forewarning in
fear arousing communication. Psychological Reports, 58, 87-90.
Gaertner, S. L., & Dovidio, J. F. (1986). The aversive form of racism. In J. F. Dovidio & S. L. Gaertner
(Eds.), Prejudice, discrimination, and racism (pp. 61-89). Orlando, FL: Academic Press.
Gerard, H. B., Blevans, S. A., & Malcolm T. (1964). Self-evaluation and the evaluation of choice
alternatives. Journal of Personality, 32, 395-410.
Gilbert, D. T. (1991). How mental systems believe. American Psychologist, 46, 107-119.
Gilbert, D. T. (this volume). Person perception. In D. Gilbert, S. Fiske, & G. Lindzey (Eds.) Handbook
of social psychology.
Gilbert, D. T., Tafarodi, R. W., & Malone, P. S. (1993). You can't not believe everything you read.
Journal of Personality and Social Psychology, 65, 221-233.
Giner-Sorolla, R., & Chaiken, S. (1996). Selective use of heuristic and systematic processing under
defense motivation. Personality and Social Psychology Bulletin.
Gleicher, G., & Petty, R. E. (1992). Expectations of reassurance influence the nature of fear-stimulated
attitude change. Journal of Experimental Social Psychology, 28, 86-100.
Glenn, N. O. (1980). Values, attitudes, and beliefs. In O. G. O’Brim & J. Kagan (Eds.), Constancy and
change in human development. Cambridge, MA: Harvard Press.
Goethals, G. R., & Nelson, R. E. (1973). Similarity in the influence process: The belief-value distinction.
Journal of Personality and Social Psychology, 25, 117-122.
Goldberg, P. (1968). Are women prejudiced against women? Transaction, 5, 28-30.
Gorn, G. J. (1982). The effects of music in advertising on choice behavior: A classical conditioning
approach. Journal of Marketing, 46, 94-101.
Gorn, G. J., & Goldberg, M. E. (1980). Children's responses to repetitive television commercials. Journal
of Consumer Research, 6, 421-424.
Gouaux, C. (1971). Induced affective states and interpersonal attraction. Journal of Personality and
Social Psychology, 20, 687-695.
Greenwald, A. G. (1968). Cognitive learning, cognitive response to persuasion, and attitude change. In
A. G. Greenwald, T. C. Brock, & T. M. Ostrom (Ed.), Psychological foundations of attitudes (pp. 147-170). New
York: Academic Press.
Greenwald, A. G., & Albert, R. D. (1968). Acceptance and recall of improvised arguments. Journal of
Personality and Social Psychology, 8, 31-34.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-esteem, and
stereotypes. Psychological Review, 102, 4-27.
Greenwald, A. G., & Ronis, D. L. (1978). Twenty years of cognitive dissonance: Case study of a theory.
Psychological Review, 85, 53-57.
Gresham, L. G., & Shimp, T. A. (1985). Attitude toward the advertisement and brand attitude: A
60
classical conditioning perspective. Journal of Advertising, 14, 10-17.
Grice, H. P. (1975). Logic and conversation. In P. Cole, & J. L. Morgan (Ed.), Syntax and semantics
(Vol. 3, pp. 41-58). New York: Academic Press.
Griffitt, W. B. (1970). Environmental effects on interpersonal affective behavior: Ambient effective
temperature and attraction. Journal of Personality and Social Psychology, 15, 240-244.
Gross, S., Holtz, R., & Miller, N. (1995). Attitude certainty. In R. E. Petty & J. A. Krosnick (Eds.).
Attitude strength: Antecedents and consequences (pp. 215-245). Mahwah, NJ: Erlbaum.
Grush, J. E. (1976). Attitude formation and mere exposure phenomena: A nonarticial explanation of
empirical findings. Journal of Personality and Social Psychology, 33, 281-290.
Hale, J., Mongeau, P.A., & Thomas, R. M. (1991). Cognitive processing of one-and two-sided persuasive
messages. Western Journal of Speech Communication, 55, 380-389.
Harasty, Petty, & Priester (1996). Self-interest versus group-interest as bases for expecting what a source
will advocate: Implications for message processing and persuasion. Unpublished manuscript, Ohio State
University, Columbus, OH.
Harkins, S. G., & Petty, R. E. (1981a). The effects of source magnification of cognitive effort on
attitudes: An information processing view. Journal of Personality and Social Psychology, 40, 401-413.
Harkins, S. G., & Petty, R. E., (1981b). The multiple source effect in persuasion: The effects of
distraction. Personality and Social Psychology Bulletin, 7, 627-635.
Harkins, S. G., & Petty, R. E. (1987). Information utility and the multiple source effect. Journal of
Personality and Social Psychology, 52, 260-268.
Harrison, A. A. (1977). Mere exposure. In L. Berkowitz (Ed.), Advances in Experimental Social
Psychology (Vol. 10, pp. 40-85).
Hartmann, G. W. (1936). A field experiment on the comparative effectiveness of <emotional’ and
<rational’ political leaflets in determining election results. Journal of Abnormal and Social Psychology, 31, 99114.
Hass, R. G. (1981). Effects of source characteristics on cognitive responses and persuasion. In R. E.
Petty, T. M. Ostrom, & T. C. Brock (Eds.), Cognitive responses in persuasion (pp. 141-172). Hillsdale, NJ:
Erlbaum.
Hass, R. G., & Grady, K. (1975). Temporal delay, type of forewarning, and resistance to influence.
Journal of Experimental Social Psychology, 11, 459-469.
Hass, R. G., & Linder, D. E. (1972). Counterargument availability and the effects of message structure
on persuasion. Journal of Personality and Social Psychology, 23, 219-233.
Hastie, R., & Park, B. (1986). The relationship between memory and judgment depends on whether the
judgment task is memory-based or on-line. Psychological Review, 93, 258-268.
Haugtvedt, C. P., & Petty, R. E. (1992). Personality and persuasion: Need for cognition moderates the
persistence and resistance of attitude changes. Journal of Personality and Social Psychology, 63, 308-319.
Haugtvedt, C. P., Petty, R. E., & Cacioppo, J. T. (1992). Need for cognition and advertising:
Understanding the role of personality variables in consumer behavior. Journal of Consumer Psychology, 1,
239-260.
Haugtvedt, C. P., Schumann, D. W., Schneier, W. L., & Warren, W. L. (1994). Advertising repetition
and variation strategies: Implications for understanding attitude strength. Journal of Consumer Research, 21, 176204.
Haugtvedt, C. P., & Wegener, D. T. (1994). Message order effects in persuasion: An attitude strength
perspective. Journal of Consumer Research, 21, 205-218.
Heesacker, M. H., Petty, R. E., & Cacioppo, J. T. (1983). Field dependence and attitude change: Source
credibility can alter persuasion by affecting message-relevant thinking. Journal of Personality, 51, 653-666.
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Henninger, M., & Wyer, R. S. (1976). The recognition and elimination of inconsistencies among
syllogistically related beliefs: Some new light on the "Socratic effect." Journal of Personality and Social
Psychology, 34, 680-693.
Higgins, E. T. (1981). The “communication game”: Implications for social cognition and persuasion. In
61
E. T. Higgins, C. P. Herman, & M. P. Zanna (Eds.). Social cognition: The Ontario symposium (Vol. 1, pp. 343392). Hillsdale, NJ: Erlbaum.
Higgins, E. T. (1989). Self-discrepancy theory: What patterns of self-beliefs cause people to suffer? In
L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 22, pp, 93-136). New York: Academic
Press.
Higgins, E. T., & McCann, C. D. (1984). Social encoding and subsequent attitudes, impressions, and
memory: "Context-driven" and motivational aspects of processing. Journal of Personality and Social Psychology,
47, 26-39.
Hilgard, E. R. (1965). Hypnotic susceptibility. New York, NY: Harcourt, Brace, and World.
Himmelfarb, S. (1972). Integration and attribution theories in personality impression formation. Journal
of Personality and Social Psychology, 23, 309-313.
Houston, D. A., & Fazio, R. H. (1989). Biased processing as a function of attitude accessibility: Making
objective judgments subjectively. Social Cognition, 7, 51-66.
Hovland, C. I. (1957). Order of presentation in persuasion. New Haven, CT: Yale University Press.
Hovland, C. I., Harvey, O. J., & Sherif, M. (1957). Assimilation and contrast in communication and
attitude change. Journal of Abnormal and Social Psychology, 55, 242-252.
Hovland, C. I., & Janis, I. L. (1959). Personality and persuasibility. New Haven, CT: Yale University
Press.
Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion: Psychological
studies of opinion change. New Haven, CT: Yale University Press.
Hovland, C. I., Lumsdaine, A. A., & Sheffield, F. D. (1949). Experiments on mass communication.
Princeton, NJ: Princeton University Press.
Hovland, C. I., & Mandell, W. (1957). Is there a <law of primacy’ in persuasion? In C. I. Hovland (Ed.),
The order of presentation in persuasion (pp. 1-22). New Haven, CT: Yale University Press.
Hovland, C. I., & Rosenberg, M. J. (1960). Attitude organization and change. New Haven, CT: Yale
University Press.
Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on communication effectiveness.
Public Opinion Quarterly, 15, 635-650.
Howard, D. J. (1990). Rhetorical question effects on message processing and persuasion: The role of
information availability and the elicitation of judgment. Journal of Experimental Social Psychology, 26, 217-239.
Hunter, J. E., Danes, J. E., & Cohen, S. H. (1984). Mathematical models of attitude change: Change in
single attitudes and cognitive structure. San Diego, CA: Academic Press.
Ickes, W., Robertson, E., Tooke, W., & Teng, G. (1986). Naturalistic social cognitions: Methodology,
assessment, and validation. Journal of Personality and Social Psychology, 51, 66-82.
Insko, C. A. (1964). Primacy versus recency in persuasion as a function of the timing of arguments and
measures. Journal of Abnormal and Social Psychology, 69, 381-391.
Insko, C. A., Songer, E., & McGarvey, W. (1974). Balance, positivity, and agreement in the Jordan
paradigm: A defense of balance theory. Journal of Experimental Social Psychology, 10, 53-83.
Insko, C. A., Turnbull, W., & Yandell, B. (1974). Facilitative and inhibiting effects of distraction on
attitude change. Sociometry, 37, 508-528.
Isen, A. M., Daubman, K. A., & Nowicki, G. P. (1987). Positive affect facilitates creative problem
solving. Journal of Personality and Social Psychology, 52, 1122-1131.
Jacoby, L. L., Kelley, C. M., Brown, J., & Jasechko, J. (1989). Becoming famous overnight: Limits on
the ability to avoid unconscious influences of the past. Journal of Personality and Social Psychology, 56, 326338.
Jamieson, D. W., & Zanna, M. P. (1989). Need for structure in attitude formation and expression. In A.
R. Pratkanis, S. J. Breckler, & A. G. Greenwald (Ed.), Attitude structure and function (pp. 383-406). Hillsdale,
NJ: Erlbaum.
Janis, I. L. (1954). Personality correlates of susceptibility to persuasion. Journal of Personality, 22, 504518.
Janis, I. L. (1967). Effects of fear arousal on attitude change: Recent developments in theory and
62
experimental research. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 3, pp. 166-224).
San Diego, CA: Academic Press.
Janis, I. L., & Feshbach, S. (1953). Effects of fear-arousing communications. Journal of Abnormal and
Social Psychology, 48, 78-92.
Janis, I. L., & Field, P. B. (1959). Sex differences and personality factors related to persuasibility. In C.
I. Hovland & I. L. Janis (Eds.), Personality and persuasibility (pp. 55-68). New Haven, CT: Yale University
Press.
Janis, I. L., & King, B. T. (1954). The influence of role playing on opinion change. Journal of Abnormal
and Social Psychology, 49, 211-218.
Jarvis, W.B.G., & Petty, R. E. (1996). The need to evaluate. Journal of Personality and Social
Psychology,
Jellison, J., & Arkin, R. (1977). Social comparison of abilities: A self-presentation approach to decision
making in groups. In J. Suls & R. Miller (Eds.), Social comparison processes: theoretical and empirical
perspectives. Washington, DC: Himisphere.
Jepson, C., & Chaiken, S. (1990). Chronic issue-specific fear inhibits systematic processing of persuasive
communications. Journal of Social Behavior and Personality, 5, 61-84.
Johnson, B. T., & Eagly, A. H. (1989). Effects of involvement on persuasion: A meta-analysis.
Psychological Bulletin, 106, 290-314.
Johnson, B.T., Lin, H. Y., Symons, C.S., Campbell, L., & Ekstein, G. (1995). Initial beliefs and attitude
latitudes as factors in persuasion. Personality and Social Psychology Bulletin, 21, 502-511.
Johnson, E., & Tversky, A. (1983). Affect, generalization, and the perception of risk. Journal of
Personality and Social Psychology, 45, 20-31.
Johnson, H. H., & Watkins, T. A. (1971). The effects of message repetition on immediate and delayed
attitude change. Psychonomic Science, 22, 101-103.
Johnson, R. W., Kelly, R. J., & LeBlanc, B. A. (1995). Motivational basis of dissonance: Aversive
consequences or inconsistency. Personality and Social Psychology Bulletin, 21, 502-511.
Jones, E. E., & Davis, K. E. (1965). From acts to dispositions: The attribution process in person
perception. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp. 219-266). New York:
Academic Press.
Jones, E. E., Rhodewalt, F., Berglas, S., & Skelton, J. A. (1981). Effects of strategic self-presentation on
subsequent self-esteem. Journal of Personality and Social Psychology, 41, 407-421.
Judd, C. M., Drake, R. A., Downing, J. W., & Krosnick, J. A. (1991). Some dynamic properties of
attitude structures: Context-induced response facilitation and polarization. Journal of Personality and Social
Psychology, 60, 193-202.
Judd, C. M., & Lusk, C. M. (1984). Knowledge structures and evaluative judgments: Effects of
structural variables on judgment extremity. Journal of Personality and Social Psychology, 46, 1193-1207.
Kahneman, D., Knetch, J., & Thaler, R. (1991). The endowment effect, los aversion, and status quo bias.
Journal of Economic Perspectives, 5, 193-206.
Kallgren, C. A., & Wood, W. (1986). Access to attitude-relevant information in memory as a
determinant of attitude-behavior consistency. Journal of Experimental Social Psychology, 22, 328-338.
Kanouse, D. E., & Hanson. L. R. (1971). Negativity in evaluations. In E. E. Jones et al., Attribution:
Perceiving the causes of behavior (pp. 47-62).. Morristown: General Learning Press.
Kaplan, K. J. (1972). On the ambivalence-indifference problem in attitude theory and measurement: A
suggested modification of the semantic differential technique. Psychological Bulletin, 77, 361-372.
Kardes, F. R. (1994). Consumer judgment and decision processes. In R. S. Wyer and T. K. Srull (Eds.),
Handbook of social cognition, 2nd Edition (Vol. 2; pp. 399-466). Hillsdale, NJ: Erlbaum.
Kassin, S. M., Reddy, M. E., & Tulloch, W. F. (1990). Juror interpretations of ambiguous evidence: The
need for cognition, presentation order, and persuasion. Law and Human Behavior, 14, 43-55.
Katz, D. (1960). The functional approach to the study of attitudes. Public Opinion Quarterly, 24,
163-204.
Kelman, H. C. (1953). Attitude change as a function of response restriction. Human Relations, 6, 185-
63
214.
Kelman, H. C. (1958). Compliance, identification, and internalization: Three processes of attitude
change. Journal of Conflict Resolution, 2, 51-60.
Kelman, H. C., & Hovland, C. I. (1953). "Reinstatement" of the communicator in delayed measurement
of opinion change. Journal of Abnormal and Social Psychology, 48, 327-335.
Kiesler, C. A., & Kiesler, S. B. (1964). Role of forewarning in persuasive communications. Journal of
Abnormal and Social Psychology, 68, 547-549.
Kiesler, C. A., & Munson, P. A. (1975). Attitudes and opinions. Annual Review of Psychology, 26, 415456.
Kiesler, S. B., & Mathog, R. (1968). The distraction hypothesis in attitude change. Psychological
Reports, 23, 1123-1133.
Knower, F. H. (1935). Experimental studies of change in attitude: I. A study of the effect of oral
arguments on changes of attitudes. Journal of Social Psychology, 6, 315-347.
Knower, F. H. (1936). Experimental studies of change in attitude: II. A study of the effect of printed
argument on changes in attitude. Journal of Abnormal and Social Psychology, 30, 522-532.
Kraus, S. J. (1995). Attitudes and the prediction of behavior: A meta-analysis of the empirical literature.
Personality and Social Psychology Bulletin, 21, 58-75.
Krosnick, J. A., & Alwin, D. F. (1989). Aging and susceptibility to attitude change. Journal of
Personality and Social Psychology, 57, 416-425.
Krosnick, J. A., Betz, A. L., Jussim, L. J., & Lynn, A. R. (1992). Subliminal conditioning of attitudes.
Personality and Social Psychology Bulletin, 18, 152-162.
Krosnick, J. A., & Petty, R. E. (1995). Attitude strength: An overview. In R. E. Petty & J. A. Krosnick
(Eds.), Attitude strength: Antecedants and consequences (pp. 1-24). Mahwah, NJ: Erlbaum.
Kruglanski, A. W. (1990). Motivations for judging and knowing: Implications for causal attribution. In
E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of motivation and cognition: Foundations of social behavior
(Vol. 2, pp. 333-368). New York, NY: Guilford Press.
Kruglanski, A. W., & Freund, T. (1983). The freezing and unfreezing of lay-inferences: Effects on
impressional primacy, ethnic stereotyping, and numerical anchoring. Journal of Experimental Social Psychology,
19, 448-468.
Kruglanski, A. W., Freund, T., & Bar-Tal, D. (1996). Motivational effects in the mere-exposure
paradigm. European Journal of Social Psychology, 26, 479-499.
Kruglanski, A. W., & Mackie, D. M. (1990). Majority and minority influence: A judgmental process
analysis. In W. Stroebe, & M. Hewstone (Ed.), European review of social psychology (Vol. 1, pp. 229-262). New
York: John Wiley.
Kunst-Wilson, W. R., & Zajonc, R. B. (1980). Affective discrimination of stimuli that cannot be
recognized. Science, 207, 557-558.
Lambert, A. J., Cronen, S., Chasten, A. L., & Lickel, B. (1996). Private vs. public expressions of racial
prejudice. Journal of Experimental Social Psychology, 32, 437-459.
Lana, R. E. (1964). Three interpretations of order effects in persuasive communications. Psychological
Bulletin, 61, 314-320.
Lassiter, G. D., Apple, K. J., & Leach, J. G. (1994). How a high need for cognition can increase or
decrease attitude polarization. Paper presented at the annual meeting of the American Psychological Society,
Washington, D. C.
Lasswell, H. D. (1948). The structure and function of communication in society. In L. Bryson (Ed.), The
communication of ideas: Religion and civilization series (pp. 37-51). New York: Harper & Row.
Lasswell, H. D., Casey, R. D., & Smith, B. L. (1935). Propaganda and promotional activities.
Minneapolis: University of Minnesota Press.
Lazarsfeld, P. F., Berelson, B., & Gaudet, H. (1944). The people’s choice. New York: Duell, Sloan, and
Pearce.
Lefebvre, V. A. (1985). The golden section and an algebraic model of ethical cognition. Journal of
Mathematical Psychology, 29, 289-310.
64
Leippe, M. R., & Elkin, R. A. (1987). When motives clash: Issue involvement and response
involvement as determinants of persuasion. Journal of Personality and Social Psychology, 52, 269-278.
Leone, C., & Ensley, E. (1986). Self-generated attitude change: A person by situation analysis of attitude
polarization and attenuation. Journal of Research in Personality, 20, 434-446.
Lepper, M. R., Greene, D., & Nisbett, R. E. (1973). Undermining children's intrinsic interest with
extrinsic reward: A test of the "overjustification" hypothesis. Journal of Personality and Social Psychology, 28,
129-137.
Levenson, H., Burford, B., & Davis, L. (1975). Are women still prejudiced against women? A
replication and extension of Goldberg’s study. Journal of Psychology, 89, 67-71.
Leventhal, H. (1970). Findings and theory in the study of fear communications. In L. Berkowitz (Ed.),
Advances in experimental social psychology (Vol. 5, pp. 119-186). Sand Diego, CA: Academic Press.
Leventhal, H., & Niles, P. (1965). Persistence of influence for varying durations of exposure to threat
stimuli. Psychological Reports, 16, 223-233.
Levine, G. L., Halberstadt, J. B., & Goldstone, R. L. (1996). Reasoning and the weighting of attributes in
attitude judgments. Journal of Personality and Social Psychology, 70, 230-240.
Levine, J. M., Saxe, L., & Harris, H. J. (1976). Reaction to opinion deviance: Impact of deviate's
direction and distance of movement. Sociometry, 39, 97-107.
Lewan, P. C., & Stotland, E. (1961). The effects of prior information on susceptibility to an emotional
appeal. Journal of Abnormal and Social Psychology, 62, 450-453.
Lord, C. G., Ross, L., & Lepper, M. R. (1979). Biased assimilation and attitude polarization: The effects
of prior theories on subsequently considered evidence. Journal of Personality and Social Psychology, 37,
2098-2109.
Losch, M. E., & Cacioppo, J. T. (1990). Cognitive dissonance may enhance sympathetic tonus, but
attitudes are changed to reduce negative affect rather than arousal. Journal of Experimental Social Psychology,
26, 289-304.
Love, R. E., & Greenwald, A. G. (1978). Cognitive responses to persuasion as mediators of opinion
change. Journal of Social Psychology, 104, 231-241.
Lumsdaine, A. A., & Janis, I. L. (1953). Resistance to <counterpropaganda’ produced by one-sided and
two-sided <propaganda’ presentations. Public Opinion Quarterly, 17, 311-318.
Lund, F. H. (1925). The psychology of belief: IV. The law of primacy in persuasion. Journal of
Abnormal and Social Psychology, 20, 183-191.
Lutz, R. J. (1975). Changing brand attitudes through modification of cognitive structure. Journal of
Consumer Research, 1, 49-59.
Lutz, R. J. (1977). An experimental investigation of causal relations among cognitions, affect, and
behavioral intention. Journal of Consumer Research, 3, 197-208.
Lydon, J., Zanna, M. P., & Ross, M. (1988). Bolstering attitudes by autobiographical recall: Attitude
persistence and selective memory. Personality and Social Psychology Bulletin, 14, 78-86.
Maass, A., & Clark, R. D. (1983). Internalization versus compliance: Differential processes underlying
minority influence and conformity. European Journal of Social Psychology, 13, 197-215.
Maass, A., & Clark, R. D. (1984). Hidden impact of minorities: Fifteen years of minority influence
research. Psychological Bulletin, 95, 428-450.
Maccoby, E. E., & Jacklin, C. N. (1974). The psychology of sex differences. Stanford, CA: Stanford
University Press.
MacKenzie, S. B. (1986). The role of attention in mediating the effect of advertising on attribute
importance. Journal of Consumer Research, 13, 174-195.
Mackie, D. M. (1987). Systematic and nonsystematic processing of majority and minority persuasive
communications. Journal of Personality and Social Psychology, 53, 41-52.
Mackie, D. M., & Asuncion, A. G. (1990). On-line and memory-based modification of attitudes:
Determinants of message recall-attitude change correspondence. Journal of Personality and Social Psychology,
59, 5-16.
Mackie, D. M., & Worth, L. T. (1989). Processing deficits and the mediation of positive affect in
65
persuasion. Journal of Personality and Social Psychology, 57, 27-40.
Mackie, D. M., & Worth, L. T. (1991). Feeling good, but not thinking straight: The impact of positive
mood on persuasion. In J. Forgas (Ed.), Emotion and social judgment (Oxford, England: Pergamon Press.
Mackie, D. M., Worth, L. T., & Asuncion, A. G. (1990). Processing of persuasive in-group messages.
Journal of Personality and Social Psychology, 58, 812-822.
Mahesharan, D., & Chaiken, S. (1991). Promoting systematic processing in low-motivation settings:
Effect of incongruent information on processing and judgment. Journal of Personality and Social Psychology, 61,
13-33.
Maheswaran, D., & Myers-Levy, J. (1990). The influence of message framing and issue involvement.
Journal of Marketing Research, 27, 361-367.
Maio, G. R., Bell, D. W., & Esses, V. M. (1996). Ambivalence and persuasion: The processing of
messages about immigrant groups. Journal of Experimental Social Psychology.
Mandler, G., Nakamura, Y., & Shebo Van Zandt, B. J. (1987). Nonspecific effects of exposure on
stimuli that cannot be recognized. Journal of Experimental Psychology: Learning, Memory, and Cognition, 13,
646-648.
Mannheim, K. (1952). The problem of generations. In P. Kecskemeti (Ed.), Essays on the sociology of
knowledge (pp. 276-322). London: Routledge & Kegan Paul.
Markus, H., & Zajonc, R. B. (1985). The cognitive perspective in social psychology. In G. Lindzey & E.
Aronson (Eds.), Handbook of social psychology (Vol. 1, pp. 137-230). Hillsdale, NJ: Erlbaum.
Martin, L. L., Seta, J. J., & Crelia, R. A. (1990). Assimilation and contrast as a function of people’s
willingness and ability to expend effort in forming an impression. Journal of Personality and Social Psychology,
59, 27-37.
Martin, L. L., & Tesser, A. (1996). Ruminative thought. In R. S. Wyer (Ed.). Advances in social
cognition.
Mathur, M., & Chattopadhyay, A. (1991). The impact of moods generated by television programs on
responses to advertising. Psychology and Marketing, 8, 59-77.
Mayer, J. D., Gaschke, Y. N., Bravermen, D. L., & Evans, T. W. (1992). Mood-congruent judgment is a
general effect. Journal of Personality and Social Psychology, 63, 119-132.
McGuire, W. J. (1960a). Cognitive consistency and attitude change. Journal of Abnormal and Social
Psychology, 60, 345-353.
McGuire, W. J. (1960b). A syllogistic analysis of cognitive relationships. In C. I. Hovland, & M. J.
Rosenberg (Ed.), Attitude organization and change: An analysis of consistency among attitude components (pp.
65-111). New Haven, CT: Yale University Press.
McGuire, W. J. (1964). Inducing resistance to persuasion: Some contemporary approaches. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 1, pp. 191-229). New York: Academic.
McGuire, W. J. (1968). Personality and attitude change: An information-processing theory. In A. G.
Greenwald, T. C. Brock, & T. M. Ostrom (Ed.), Psychological foundations of attitudes (pp. 171-196). New York:
Academic.
McGuire, W. J. (1969). The nature of attitudes and attitude change. In G. Lindzey, & E. Aronson (Ed.),
Handbook of social psychology (Vol. 3, pp. 136-314). Reading, MA: Addison-Wesley.
McGuire, W. J. (1981). The probabilogical model of cognitive structure and attitude change. In R. E.
Petty, T. M. Ostrom, & T. C. Brock (Ed.), Cognitive responses in persuasion (pp. 291-307). Hillsdale, NJ:
Erlbaum.
McGuire, W. J. (1985). Attitudes and attitude change. In G. Lindzey, & E. Aronson (Ed.), Handbook of
social psychology (Vol. 2, pp. 233-346). New York: Random House.
McGuire, W. J. (1989). Theoretical foundations of campaigns. In R. E. Rice, & C. K. Atkin (Ed.), Public
communication campaigns (pp. 43-66). Newbury Park: Sage.
McGuire, W. J. (1995a). The communication and attitude change program at Yale in the 1950s. In E. E.
Dennis & E. Wartella (Eds.), American communication research, the remembered history. Hillsdale, NJ:
Erlbaum.
McGuire, W. J. (1995b). Transferring research findings on persuasion to improve drug abuse prevention
66
programs. In T. E. Backer, S. L. David, & G. Soucy (Eds.), Reviewing the behavioral science knowledge base on
technology transfer (pp. 225-245). Rockville, MD: National Institute on Drug Abuse (NIH Publication No. 954035).
McGuire, W. J., & McGuire, C. V. (1991). The content, structure, and operation of thought systems. In
R. S. Wyer Jr., & T. Srull (Ed.), Advances in social cognition (Vol. 4, pp. 1-78). Hillsdale, NJ: Erlbaum.
McGuire, W.J., & McGuire, C. V. (1996). Enhancing self-esteem by directed-thinking tasks: Cognitive
and affective positivity asymmetries. Journal of Personality and Social Psychology, 70, 1117-1125.
McGuire, W. J., & Papageorgis, D. (1961). The relative efficacy of various types of prior belief-defense
in producing immunity against persuasion. Journal of Abnormal and Social Psychology, 62, 327-337.
McGuire, W. J., & Papageorgis, D. (1962). Effectiveness of forewarning in developing resistance to
persuasion. Public Opinion Quarterly, 26, 24-34.
Messerschmidt, R. (1933). The suggestibility of boys and girls between the ages of six and sixteen years.
Journal of Genetic Psychology, 43, 422-427.
Meyerowitz, B. E., & Chaiken, S. (1987). The effect of message framing on breast self-examination
attitudes, intentions, and behavior. Journal of Personality and Social Psychology, 52, 500-510.
Millar, M. G., & Millar, K. U. (1990). Attitude change as a function of attitude type and argument type.
Journal of Personality and Social Psychology, 59, 217-228.
Millar, M. G., & Tesser, A. (1992). The role of beliefs and feelings in guiding behavior: The mis-match
model. In C. Martin, & A. Tesser (Ed.), Construction of social judgment (pp. 277-300). Hillsdale, NJ: Erlbaum.
Miller, A. G., McHoskey, J. W., Bane, C. M., & Dowd, T. G. (1993). The attitude polarization
phenomenon: Role of response measure, attitude extremity, and behavioral consequences of reported attitude
change. Journal of Personality and Social Psychology, 64, 561-574.
Miller, N., & Baron, R. S. (1968). Distraction, communicator credibility and attitude change.
Unpublished manuscript, University of Iowa, Iowa City, IA.
Miller, N., & Campbell, D. T. (1959). Recency and primacy in persuasion as a function of the timing of
speeches and measurements. Journal of Abnormal and Social Psychology, 59, 1-9.
Miller, N., Maruyama, G., Beaber, R., & Valone, K. (1976). Speed of speech and persuasion. Journal of
Personality and Social Psychology, 34, 615-625.
Miller, C. E., & Norman, R. M. G. (1976). Balance, agreement, and attraction in hypothetical social
situations. Journal of Experimental Social Psychology, 12, 109-119.
Mills, J., & Harvey, J. H. (1972). Opinion change as a function of when information about the
communicator is received and whether he is attractive or expert. Journal of Personality and Social Psychology,
21, 52-55.
Mills, J., & Jellison, J. M. (1967). Effect on opinion change of how desirable the communication is to
the audience the communicator addressed. Journal of Personality and Social Psychology, 6, 98-101.
Mitnick, L., & McGinnies, E. (1958). Influencing ethnocentrism in small discussion groups through a
film communication. Journal of Abnormal and Social Psychology, 56, 82-92.
Moore, D. L., Hausknecht, D., & Thamodaran, K. (1986). Time compression, response opportunity, and
persuasion. Journal of Consumer Research, 13, 85-99.
Moore, D. L., & Reardon, R. (1987). Source magnification: The role of multiple sources in the
processing of advertising appeals. Journal of Marketing Research, 24, 412-417.
Moscovici, S. (1980). Toward a theory of conversion behavior. In L. Berkowitz (Ed.), Advances in
experimental social psychology (Vol. 13, pp. 209-239). New York: Academic Press.
Moscovici, S., Lage, E., & Naffrechoux, M. (1969). Influence of a consistent minority on the responses
of a majority in a color perception task. Sociometry, 32, 365-380.
Moscovici, S., & Neve, P. (1973). Studies in social influence: II. Instrumental and symbolic influence.
European Journal of Social Psychology, 3, 461-471.
Mugny, G., & Perez, J. A. (1991). The social psychology of minority influence. Translated by V. W.
Lamongie . New York: Cambridge University Press.
Mullis, J. P., & Lippa, R. (1990). Behavioral change in earthquake preparedness due to negative threat
appeals: A test of protection motivation theory. Journal of Applied Social Psychology, 20, 619-638.
67
Murchison, C. (1935). A handbook of social psychology. Worcester, MA: Clark University Press.
Murphy, S. T., Monahan, J. L., & Zajonc, R. B. (1995). Additivity of nonconscious affect: Combined
effects of priming and exposure. Journal of Personality and Social Psychology, 69, 589-602.
Murphy, S. T., & Zajonc, R. B. (1993). Affect, cognition, and awareness: Affective priming with
optimal and suboptimal exposures. Journal of Personality and Social Psychology, 64, 723-739.
Murray, N., Sujan, H., Hirt, E. R., & Sujan, M. (1990). The influence of mood on categorization: A
cognitive flexibility interpretation. Journal of Personality and Social Psychology, 59, 411-425.
Nemeth, C. J., (1986). Differential contributions of majority and minority influence. Psychological
Review, 93, 10-20.
Newcomb, T. M. (1956). The prediction of interpersonal attraction. American Psychologist, 11, 575586.
Newcomb, T. M. (1968). Interpersonal balance. In R. P. Abelson, E. Aronson, W. J. McGuire, T. M.
Newcomb, M. J. Rosenberg, & P. H. Tannenbaum (Eds.), Theories of cognitive consistency: A sourcebook (pp.
28-51). Chicago, IL: Rand McNally.
Newcombe, N., & Arnkoff, D. B. (1979). Effects of speech style and sex of speaker on person
perception. Journal of Personality and Social Psychology, 37, 1293-1303.
Nisbett, R. E., & Gordon, A. (1967). Self-esteem and susceptibility to social influence. Journal of
Personality and Social Psychology, 5, 268-276.
Norman, R. (1975). Affective-cognitive consistency, attitudes, conformity, and behavior. Journal of
Personality and Social Psychology, 32, 83-91.
Nuttin, J. M., Jr. (1985). Narcissism beyond Gestalt and awareness: The name letter effect. European
Journal of Social Psychology, 15, 353-361.
O’Keefe, D. J. (1990). Persuasion: Theory and research. Newbury Park, CA: Sage.
Olson, J. M., & Zanna, M. P. (1993). Attitude and attitude change. Annual Review of Psychology.
Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement of meaning. Urbana, IL:
University of Illinois Press.
Osgood, C. E., & Tannenbaum, P. H. (1955). The principle of congruity in the prediction of attitude
change. Psychological Review, 62, 42-55.
Osterhouse, R. A., & Brock, T. C. (1970). Distraction increases yielding to propaganda by inhibiting
counterarguing. Journal of Personality and Social Psychology, 15, 344-358.
Pallak, S. R. (1983). Salience of a communicator's physical attractiveness and persuasion: A heuristic
versus systematic processing interpretation. Social Cognition, 2, 158-170.
Papageorgis, D. (1968). Warning and persuasion. Psychological Bulletin, 70, 271-282.
Pechman, C., & Estaban, G. (1994). Persuasion processes associated with direct comparative and
noncomparative advertising and implications for advertising effectiveness. Journal of Consumer Psychology, 2,
403-432.
Peeters, G., & Czapinski, J. (1990). Positive-negative asymmetry in evaluations: The distinction between
affective and informational negativity effects. European Review of Social Psychologoy, 1, 33-60.
Perloff, R. M., & Brock, T. C. (1980). And thinking makes it so: Cognitive responses to persuasion. In
M. Roloff, & G. Miller (Ed.), Persuasion: New directions in theory and research (pp. 67-100). Beverly Hills:
Sage.
Peterson, R. C., & Thurstone, L. L. (1933). The effect of motion pictures on the social attitudes of high
school children. Chicago: University of Chicago Press.
Petty, R. E. (1977). A cognitive response analysis of the temporal persistence of attitude changes
induced by persuasive communications. Doctoral dissertation, Ohio State University, Columbus, OH.
Petty, R. E. (1994). Two routes to persuasion: State of the art. In P. Eelen, M. Pichelle, G. d'Ydewalle,
& P. Bertelson (Eds.), 25th International Congress of Psychology: State of the art lectures. Hillsdale, NJ:
Erlbaum.
Petty, R. E. (1995). Attitude change. In A. Tesser (Ed.). Advanced social psychology (pp. 195-255).
New York: McGraw Hill.
Petty, R. E. (1996). The evolution of theory and research in social psychology: From single to multiple
68
effect and process models of persuasion. In C. McGarty & S. A. Haslam (Eds.), The message of social
psychology: Perspectives on mind in society. Oxford, England: Basil Blackwell, Ltd.
Petty, R. E., & Brock, T. C. (1976). Effects of responding or not responding to hecklers on audience
agreement with a speaker. Journal of Applied Social Psychology, 6, 1-17.
Petty, R. E., & Cacioppo, J. T. (1977). Forewarning, cognitive responding, and resistance to persuasion.
Journal of Personality and Social Psychology, 35, 645-655.
Petty, R. E., & Cacioppo, J. T. (1979a). Effects of forewarning of persuasive intent and involvement on
cognitive responses. Personality and Social Psychology Bulletin, 5, 173-176.
Petty, R. E., & Cacioppo, J. T. (1979b). Issue-involvement can increase or decrease persuasion by
enhancing message-relevant cognitive responses. Journal of Personality and Social Psychology, 37, 1915-1926.
Petty, R. E., & Cacioppo, J. T. (1981). Attitudes and persuasion: Classic and contemporary approaches.
Dubuque, IA: Wm. C. Brown.
Petty, R. E., & Cacioppo, J. T. (1983). The role of bodily responses in attitude measurement and change.
In J. T. Cacioppo, & R. E. Petty (Eds.), Social psychophysiology: A sourcebook (pp. 51-101). New York:
Guilford.
Petty, R. E., & Cacioppo, J. T. (1984). The effects of involvement on responses to argument quantity and
quality: Central and peripheral routes to persuasion. Journal of Personality and Social Psychology, 46, 69-81.
Petty, R. E., & Cacioppo, J. T. (1986a). Communication and persuasion: Central and peripheral routes
to persuasion. New York: Springer-Verlag.
Petty, R. E., & Cacioppo, J. T. (1986b). The Elaboration Likelihood Model of persuasion. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 123-205). New York: Academic
Press.
Petty, R. E., & Cacioppo, J. T. (1990). Involvement and persuasion: Tradition versus integration.
Psychological Bulletin, 107, 367-374.
Petty, R. E., & Cacioppo, J. T. (1996). Addressing disturbing and disturbed consumer behavior: Is it
necessary to change the way we conduct behavioral science? Journal of Marketing Research.
Petty, R. E., Cacioppo, J. T., & Goldman, R. (1981). Personal involvement as a determinant of
argument-based persuasion. Journal of Personality and Social Psychology, 41, 847-855.
Petty, R. E., Cacioppo, J. T., & Haugtvedt, C. (1992). Involvement and persuasion: An appreciative
look at the Sherifs' contribution to the study of self-relevance and attitude change. In D. Granberg, & G. Sarup
(Ed.), A Social judgment and intergroup relations: Essays in honor of Muzifer Sherif (pp. 147-175). New York:
Springer-Verlag.
Petty, R. E., Cacioppo, J. T., & Heesacker, M. (1981). Effects of rhetorical questions on persuasion: A
cognitive response analysis. Journal of Personality and Social Psychology, 40, 432-440.
Petty, R. E., Cacioppo, J. T., & Schumann, D. W. (1983). Central and peripheral routes to advertising
effectiveness: The moderating role of involvement. Journal of Consumer Research, 10, 135-146.
Petty, R. E., Gleicher, F., & Baker, S. M. (1991). Multiple roles for affect in persuasion. In J. Forgas
(Ed.), Emotion and social judgments (pp. 181-200). Oxford, England: Pergamon Press.
Petty, R. E., Gleicher, F., & Jarvis, W. B. G. (1993). Persuasion theory and AIDS prevention. In J. B.
Pryor, & G. D. Reeder (Eds.), The social psychology of HIV infection (pp. 155-182). Hillsdale, NJ: Erlbaum.
Petty, R. E., Harkins, S. G., & Williams, K. D. (1980). The effects of group diffusion of cognitive effort
on attitudes: An information processing view. Journal of Personality and Social Psychology, 38, 81-92.
Petty, R. E., Haugtvedt, C. P., & Smith, S.M. (1995). Elaboration as a determinant of attitude strength. In
R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and consequences (pp. 93-130). Mahwah,
NJ: Erlbaum.
Petty, R. E., & Jarvis, B. G. (1996) An individual differences perspective on assessing cognitive
processes. In N. Schwarz & S. Sudman (Eds.), Answering questions: Methodology for determining cognitive and
communicative procesess in survey research (pp. 221-257). San Francisco: Jossey-Bass.
Petty, R.E., Jarvis, B. G., & Evans, L. M. (1996). Recurrent thought: Some implications for attitudes and
persuasion. Advances in social cognition, 9.
Petty, R. E., Kasmer, J. A., Haugtvedt, C. P., & Cacioppo, J. T. (1987). A reply to Stiff and Boster.
69
Communication Monographs, 54, 257-263.
Petty, R. E., & Krosnick, J. A. (Eds.) (1995). Attitude strength: Antecedents and consequences.
Mahwah, NJ: Erlbaum.
Petty, R. E., Ostrom, T. M., & Brock, T. C. (Eds.) (1981). Cognitive responses in persuasion. Hillsdale,
NJ: Erlbaum.
Petty, R. E., Priester, J. R., & Wegener, D. T. (1994). Cognitive processes in attitude change. In R. S.
Wyer and T. K. Srull (Eds.) Handbook of social cognition (2nd edition), (Vol. 2, pp. 69-142). Hillsdale, NJ:
Erlbaum.
Petty, R. E., Schumann, D. W., Richman, S. A., & Strathman, A. J. (1993). Positive mood and
persuasion: Different roles for affect under high- and low-elaboration conditions. Journal of Personality and
Social Psychology, 64, 5-20.
Petty, R. E., & Wegener, D. T. (1991). Thought systems, argument quality, and persuasion. In R. S.
Wyer Jr., & T. K. Srull (Ed.), Advances in social cognition (Vol. 4, pp. 147-161). Hillsdale, NJ: Erlbaum.
Petty, R. E., & Wegener, D. T. (1993). Flexible correction processes in social judgment: Correcting for
context induced contrast. Journal of Experimental Social Psychology, 29, 137-165.
Petty, R. E., & Wegener, D. T. (1995, September). Correction processes in social judgment: Implications
for persuasion. Paper presented at the annual meeting of the Society of Experimental Social Psychology,
Washington, DC.
Petty, R. E., & Wegener, D. T. (1996). Match versus mismatch of persuasive appeals to functional bases
of attitudes: Effects on processing of message content. Unpublished manuscript, Columbus, OH.
Petty, R. E., Wegener, D. T., & Fabrigar, L. R. (1997). Attitudes and attitude change. Annual Review of
Psychology.
Petty, R. E., Wegener, D. T., Fabrigar, L. R., Priester, J. R., & Cacioppo, J. T. (1993). Conceptual and
methodological issues in the elaboration likelihood model of persuasion: A reply to the Michigan State critics.
Communication Theory, 3, 336-363.
Petty, R. E., Wells, G. L., & Brock, T. C. (1976). Distraction can enhance or reduce yielding to
propaganda: Thought disruption versus effort justification. Journal of Personality and Social Psychology, 34,
874-884.
Petty, R. E., Wells, G. L., Heesacker, M., Brock, T. C., & Cacioppo, J. T. (1983). The effects of
recipient posture on persuasion: A cognitive response analysis. Personality and Social Psychology Bulletin, 9,
209-222.
Pfau, M., & Kenski, H. C., Nitz, M., & Sorenson, J. (1990). Efficacy of inoculation strategies in
promoting resistance to political attack messages: Application to direct mail. Communication Monographs, 57,
25-43.
Pittman, T. S. (1994). Control motivation and attitude change. In G. Weary, F. Gleicher, & K. Marsh
(Ed.), Control Motivation and Social Cognition. New York: Springer Verlag.
Pomerantz, E. M., Chaiken, S., & Tordesillas, R. S. (1995). Attitude strength and resistance processes.
Journal of Personality and Social Psychology, 69, 408-419.
Pratkanis, A. R., Breckler, S. J., & Greenwald, A. G. (Eds.) (1989). Attitude structure and function.
Hillsdale, NJ: Erlbaum.
Pratkanis, A. R., Greenwald, A. G., Leippe, M. R., & Baumgardner, M. H. (1988). In search of reliable
persuasion effects: III. The sleeper effect is dead. Long live the sleeper effect. Journal of Personality and Social
Psychology, 54, 203-218.
Priester, J. R., Cacioppo, J. T., & Petty, R. E. (1996). The influence of motor processes on attitudes
toward novel versus familiar semantic stimuli. Personality and Social Psychology Bulletin.
Priester, J. R., & Petty, R. E. (1995). Source attributions and persuasion: Perceived honesty as a
determinant of message scrutiny. Personality and Social Psychology Bulletin, 21, 637-654.
Priester, J. R., & Petty, R. E. (1996). The gradual threshold model of ambivalence: Relating the positive
and negative bases of attitudes to subjective ambivalence. Journal of Personality and Social Psychology.
Puckett, J. M., Petty, R. E., Cacioppo, J. T., & Fisher, D. L. (1983). The relative impact of age and
attractiveness stereotypes on persuasion. Journal of Gerontology, 38, 340-343.
70
Ratneshwar, S., & Chaiken, S. (1991). Comprehension's role in persuasion: The case of its moderating
effect on the persuasive impact of source cues. Journal of Consumer Psychology, 18, 52-62.
Raven, B. H., & French, J. R. P. (1958). Legitimate power, coercive power, and observability in social
influence. Sociometry, 21, 83-97.
Razran, G. H. S. (1940). Conditioned response changes in rating and appraising sociopolitical slogans.
Psychological Bulletin, 37, 481.
Rhine, R., & Severance, L. (1970). Ego-involvement, discrepancy, source credibility, and attitude
change. Journal of Personality and Social Psychology, 16, 175-190.
Rhodes, N., & Wood, W. (1992). Self-esteem and intelligence affect influenceability: The mediating role
of message reception. Psychological Bulletin, 111, 156-171.
Rippetoe, P. A., & Rogers, R. W. (1987). Effects of components of protection-motivation theory on
adaptive and maladaptive coping with a health threat. Journal of Personality and Social Psychology, 52, 596-604.
Rogers, R. W. (1975). A protection motivation theory of fear appeals and attitude change. Journal of
Psychology, 91, 93-114.
Rogers, R. W. (1983). Cognitive and physiological processes in fear appeals and attitude change: A
revised theory of protection motivation. In J. T. Cacioppo, & R. E. Petty (Ed.), Social psychophysiology: A
sourcebook (pp. 153-176). New York: Guilford.
Romero, A. A., Agnew, C. R., & Insko, C. A. (1996). The cognitive mediation hypothesis revisited: An
empirical response to methodological and theoretical criticism. Personality and Social Psychology Bulletin, 22,
651-665.
Ronis, D. L., Baumgardner, M. H., Leippe, M. R., Cacioppo, J. T., & Greenwald, A. G. (1977). In search
of reliable persuasion effects: I. A computer-controlled procedure for studying persuasion. Journal of
Personality and Social Psychology, 35, 548-569.
Roselli, F., Skelly, J. J., & Mackie, D. M. (1995). Processing rational and emotional messages: The
cognitive and affective mediation of persuasion. Journal of Experimental Social Psychology, 31, 163-190.
Rosenbaum, M. E. (1986). The repulsion hypothesis: On the nondevelopment of relationships. Journal of
Personality and Social Psychology, 51, 1156-1166.
Rosenberg, M. J. (1956). Cognitive structure and attitudinal affect. Journal of Abnormal and Social
Psychology, 53, 367-372.
Rosenberg, M. J. (1960). Cognitive reorganization in response to hypnotic reversal of attitudinal affect.
Journal of Personality, 28, 39-63.
Rosenberg, M. J., & Abelson, R. P. (1960). An analysis of cognitive balancing. In C. I. Hovland & M. J.
Rosenberg (Eds.), Attitude organization and change: An analysis of consistency among attitude components (pp.
112-163). New Haven, CT: Yale University Press.
Roskos-Ewoldsen, D. R., & Fazio, R. H. (1992). The accessibility of source likability as a determinant of
persuasion. Personality and Social Psychology Bulletin, 18, 19-25.
Ross, M., McFarland, C., Conway, M., & Zanna, M. P. (1983). Reciprocal relation between attitudes and
behavior recall: Committing people to newly formed attitudes. Journal of Personality and Social Psychology, 45,
257-267.
Rothman, A. J., & Salovey, P. (in press). Shaping perceptions to motivate healthy behavior: The role of
message framing. Psychological Bulletin,
Rothman, A. J., Salovey, P., Antone, C., Keough, K., & Martin, C. D. (1993). The influence of message
framing on intentions to perform health behaviors. Journal of Experimental Social Psychology, 29, 408-433.
Sadler, O., & Tesser, A. (1973). Some effects of salience and time upon interpersonal hostility and
attraction. Sociometry, 36, 99-112.
Salovey, P., Rothman, A. J., & Rodin, J. (this volume). Health psychology. In D. Gilbert, S. Fiske, & G.
Lindzey (Eds.), The handbook of social psychology (4th edition).
Sanbonmatsu, D. M., & Kardes, F. R. (1988). The effects of physiological arousal on information
processing and persuasion. Journal of Consumer Research, 15, 379-385.
Scher, S. J., & Cooper, J. (1989). Motivational basis of dissonance: The singular role of behavioral
consequences. Journal of Personality and Social Psychology, 56, 899-906.
71
Schul, Y. & Burnstein, E. (1985). When discounting fails: Conditions under which individuals use
discredited information in making a judgment. Journal of Personality and Social Psychology, 49, 894-903.
Schutte, R. A., & Fazio, R. H. (1995). Attitude accessibility and motivation as determinants of biased
processing: A test of the MODE model. Personality and Social Psychology Bulletin, 21, 704-710.
Schwarz, N. (1990). Feelings as information: Informational and motivational functions of affective
states. In E. T. Higgins, & R. M. Sorrentino (Ed.), Handbook of motivation and cognition: Foundations of social
behavior (Vol. 2, pp. 527-561). New York: Guilford.
Schwarz, N., & Bless, H. (1992). Constructing reality and its alternatives: An inclusion/exclusion model
of assimilation and contrast effects in social judgment. In L. L. Martin & A. Tesser (Eds.), The construction of
social judgments (pp. 217-245). Hillsdale, NJ: Erlbaum.
Schwarz, N., Bless, H., & Bohner, G. (1991). Mood and persuasion: Affective states influence the
processing of persuasive communications. In M. P. Zanna (Ed.), Advances in experimental social psychology
(Vol. 24, pp. 161-201). San Diego: Academic Press.
Schwarz, N., & Clore, G. (1983). Mood, misattribution, and judgments of well-being: Informative and
directive functions of affective states. Journal of Personality and Social Psychology, 45, 513-523.
Scott, C. A., & Yalch, R. J. (1978). A test of the self-perception explanation of the effects of rewards on
intrinsic interest. Journal of Experimental Social Psychology, 14, 180-192.
Sears, D. O. (1981). Life stage effects on attitude change, especially among the elderly. In S. B. Kiesler,
J. N. Morgan, & V. K. Oppenheimer (Eds.), Aging: Social change (pp. 183-204). San Diego, CA: Academic
Press.
Sears, D. O. (1986). College students in the laboratory: Influences of a narrow data base on social
psychology’s view of human nature. Journal of Personality and Social Psychology, 51, 515-530.
Shavitt, S. (1989). Operationalizing functional theories of attitude. In A. R. Pratkanis, S. J. Breckler, &
A. G. Greenwald (Ed.), Attitude structure and function (pp. 311-338). Hillsdale, NJ: Erlbaum.
Shavitt, S. , Swan, S., Lowery, T. M., & Wänke, M. (1994). The interaction of endorser attractiveness
and involvement in persuasion depends on the goal that guides message processing. Journal of Consumer
Psychology, 3, 137-162.
Sherif, C. W., Sherif, M., & Nebergall, R. E. (1965). Attitude and attitude change: The social
judgment-involvement approach. Philadelphia: Saunders.
Sherif, M. (1977). Crisis in social psychology: Some remarks towards breaking through the crisis.
Personality and Social Psychology Bulletin, 3, 368-382.
Sherif, M., & Hovland, C. I. (1961). Social judgment: Assimilation and contrast effects in
communication and attitude change. New Haven, CT: Yale University Press.
Sherman, S. J. (1973). Internal-external control and its relationship to attitude change under different
social influence techniques. Journal of Personality and Social Psychology, 26, 23-29.
Sherman, S. J., Cialdini, R. B., Schwartzman, D. F., & Reynolds, K. D. (1985). Imagining can heighten
or lower the perceived likelihood of contracting a disease: The mediating effect of ease of imagery. Personality
and Social Psychology Bulletin, 11, 118-127.
Sherman, S. J., Zehner, K. S., Johnson, J., & Hirt, E. R. (1983). Social explanation: The role of timing,
set, and recall on subjective likelihood estimates. Journal of Personality and Social Psychology, 44, 1127-1143.
Shimp, T. A., Stuart, E. W., & Engle, R. W. (1991). A program of classical conditioning experiments
testing variations in the conditioned stimulus and context. Journal of Consumer Research, 18, 1-12.
Simon, L., Greenberg, J., & Brehm, J. (1995). Trivialization: The forgotten mode of dissonance
reduction. Journal of Personality and Social Psychology, 68, 247-260.
Sistrunk, F., & McDavid, J. W. (1971). Sex variable in conforming behavior. Journal of Personality and
Social Psychology, 17, 200-207.
Sivacek, J., & Crano, W. D. (1982). Vested interest as a moderator of attitude-behavior consistency.
Journal of Personality and Social Psychology, 43, 210-221.
Skolnick, P., & Heslin, R. (1971). Quality versus difficulty: Alternative interpretations of the
relationship between self-esteem and persuasibility. Journal of Personality, 39, 242-251.
Skowronski, J. J., & Carlston, D. E. (1989). Negativity and extremity biases in impression formation: A
72
review of explanations. Psychological Bulletin, 105, 131-142.
Slamecka, N. J., & Graf, P. (1978). The generation effect: Delineation of a phenomenon. Journal of
Experimental Psychology: Human Learning and Memory, 4, 592-604.
Slusher, M. P., & Anderson, C. A. (1996). Using causal persuasive arguments to change beliefs and
teach new information. The mediating role of explanation availabilitiy and evaluation bias in the acceptance of
knowledge. Journal of Educational Psychology, 88, 110-122.
Smith, M. B., Bruner, J. S., & White, R. W. (1956). Opinions and personality. New York: Wiley.
Smith, S. M., Haugtvedt, C. P., & Petty, R. E. (1994). Need for cognition and the effects of repeated
expression on attitude accessibility and extremity. Advances in Consumer Research, 21, 234-237.
Smith, S. M., & Petty, R. E. (1996). Message framing and persuasion: A message processing analysis.
Personality and Social Psychology Bulletin, 22, 257-268.
Smith, S. M., & Shaffer, D. R. (1995). Speed of speech and persuasion: Evidence for multiple effects.
Personality and Social Psychology Bulletin, 21, 1051-1060.
Snyder, M. (1974). The self-monitoring of expressive behavior. Journal of Personality and Social
Psychology, 30, 526-537.
Snyder, M., & DeBono, K. G. (1989). Understanding the functions of attitudes: Lessons from
personality and social behavior. In S. J. Pratkanis, S. J. Breckler, & A. G. Greenwald (Ed.), Attitude structure and
function (pp. 339-359). Hillsdale, N.J.: Erlbaum.
Snyder, M., & Rothbart, M. (1971). Communicator attractiveness and opinion change. Canadian Journal
of Behavioural Science, 3, 377-387.
Snyder, M., & Swann, W. B. (1976). When actions reflect attitudes: The politics of impression
management. Journal of Personality and Social Psychology, 34, 1034-1042.
Snyder, M., & Tanke, E. D. (1976). Behavior and attitude: Some people are more consistent than others.
Journal of Personality, 44, 510-517.
Sorrentino, R. M., Bobocel, D. R., Gitta, M. Z., Olson, J. M., & Hewitt, E. C. (1988). Uncertainty
orientation and persuasion: Individual differences in the effects of personal relevance on social judgments.
Journal of Personality and Social Psychology, 55, 357-371.
Sponberg, H. (1946). A study of the relative effectiveness of climax and anti-climax order in an
argumentative speech. Speech Monographs, 13, 35-44.
Staats, A. W., & Staats, C. K. (1958). Attitudes established by classical conditioning. Journal of
Abnormal and Social Psychology, 57, 37-40.
Staats, A. W., Staats, C. K., & Crawford, H. L. (1962). First-order conditioning of meaning and the
parallel conditioning of a GSR. Journal of General Psychology, 67, 159-167.
Stayman, D. M., & Kardes, F. R. (1992). Spontaneous inference processes in advertising: Effects of need
for cognition and self-monitoring on inference generation and utilization. Journal of Consumer Psychology, 1,
125-142.
Steele, C. M. (1988). The psychology of self-affirmation: Sustaining the integrity of the self. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 21, pp. 261-302). New York: Academic
Press.
Steele, C. M., Spencer, S. J., & Lynch, M. (1993). Self-image resilience and dissonance: The role of
affirmational resources. Journal of Personality and Social Psychology, 64, 885-896.
Sternthal, B., Dholakia, R., & Leavitt, C. (1978). The persuasive effect of source credibility: A test of
cognitive response analysis. Journal of Consumer Research, 4, 252-260.
Stiff, J. B. (1986). Cognitive processing of persuasive message cues: A meta analytic review of the
effects of supporting information on attitudes. Communication Monographs, 53, 75-89.
Stone, J. Aronson, E., Craine, A. L., Winslow, M. P., & Fried, C. B. (1994). Inducing hypocrisy as a
means of encouraging young adults to use condoms. Personality and Social Psychology Bulletin, 20, 116-128.
Stone, J., & Cooper, J. (1996). A self-standards model of dissonance. Unpublished manuscript.
Princeton University, Princeton, NJ.
Strack, F., Martin, L., & Stepper, S. (1988). Inhibiting and facilitating conditions of the human smile: A
nonobtrusive test of the facial feedback hypothesis. Journal of Personality and Social Psychology, 54, 768-777.
73
Stuart, E. W., Shimp, T. A., & Engle, R. W. (1987). Classical conditioning of consumer attitudes: Four
experiments in an advertising context. Journal of Consumer Research, 14, 334-349.
Sturges, J. W., & Rogers, R. R. (1996). Preventive health psychology from a developmental perspective:
An extension of protection motivation theory. Health Psychology, 15, 158-166.
Sutton, S. R. (1982). Fear-arousing communications: A critical examination of theory and research. In J.
R. Eiser (Ed.), Social psychology and behavioral medicine ( pp. 303-337). Chichester, England: Wiley.
Swim, J., Borgida, E., Maruyama, G., & Myers, D. G. (1989). Joan McKay versus John McKay: Do
gender stereotypes bias evaluations? Psychological Bulletin, 105, 409-429.
Tajfel, H. (1970). Experiments in intergroup discrimination. Scientific American, 223, 96-102.
Taylor, S. E. (1975). On inferring one's attitude from one's behavior: Some delimiting conditions.
Journal of Personality and Social Psychology, 31, 126-131.
Taylor, S. E. (1981). The interface of cognitive and social psychology. In J. H. Harvey (Ed.), Cognition,
social behavior, and the environment. Hillsdale, NJ: Erlbaum.
Tedeschi, J. T., Schlenker, B. R., & Bonoma, T. V. (1971). Cognitive dissonance: Private ratiocination
or public spectacle? American Psychologist, 26, 685-695.
Tesser, A. (1978). Self-generated attitude change. In L. Berkowitz (Ed.), Advances in experimental
social psychology (Vol. 11, pp. 289-338). New York: Academic Press.
Tesser, A. (1993). The importance of heritability in psychological research: The case of attitudes.
Psychological Review, 100, 129-142.
Tesser, A., & Cornell, D. P. (1991). On the confluence of self processes. Journal of Experimental Social
Psychology, 27, 501-526.
Tesser, A., Martin, L., & Mendolia, M. (1995). The impact of thought on attitude extremity and
attitude-behavior consistency. In R. E. Petty, & J. A. Krosnick (Ed.), Attitude strength: Antecedents and
consequences (pp. 73-92). Mahwah, NJ: Erlbaum.
Tesser, A., & Rosen, S. (1978). The reluctance to transmit bad news: The MUM effect. Advances in
Experimental Social Psychology, 8.
Tetlock, P. E. (1983). Accountability and the complexity of thought. Journal of Personality and Social
Psychology, 45, 74-83.
Tetlock, P. E. (1992). The impact of accountability on judgment and choice: Toward a social
contingency model. Advances in experimental social psychology, 25, 331-376.
Tetlock, P. E., Skitka, L., & Boettger, R. (1989). Social and cognitive strategies of coping with
accountability: Conformity, complexity, and bolstering. Journal of Personality and Social Psychology, 57, 632641.
Thibodeau, R., & Aronson, E. (1992). Taking a closer look: Reasserting the role of the self-concept in
dissonance theory. Personality and Social Psychology Bulletin, 18, 591-602.
Thompson, M. M., Zanna, M. P., & Griffin, D. W. (1995). Let's not be indifferent about (attitudinal)
ambivalence. In R. E. Petty, & J. A. Krosnick (Ed.), Attitude strength: Antecedents and consequences (pp. 361386). Mahwah, NJ: Erlbaum.
Thomsen, C. J., Borgida, E., & Lavine, H. (1995). The causes and consequences of personal
involvement. In R. E. Petty, & J. A. Krosnick (Ed.), Attitude strength: Antecedents and consequences (pp. 191214). Mahwah, NJ: Erlbaum.
Thorndike, E. L. (1920). A constant error in psychological ratings. Journal of Applied Psychology, 4, 2529.
Titchener, E. B. (1910). Textbook of psychology. New York: Macmillan.
Trost, M. R., Maass, A., & Kenrick, D. T. (1992). Minority influence: Personal relevance biases
cognitive processes and reverses private acceptance. Journal of Experimental Social Psychology, 28, 234-254.
Tykocinski, O., Higgins, E. T., & Chaiken, S. (1994). Message framing, self-discrepancies, and yielding
to persuasive messages: The motivational significance of psychological situations. Personality and Social
Psychology Bulletin, 20, 107-115.
Tyler, T. R., & Schuller, R. A. (1991). Aging and attitude change. Journal of Personality and Social
Psychology, 61, 689-697.
74
Valins, S. (1966). Cognitive effects of false heart-rate feedback. Journal of Personality and Social
Psychology, 4, 400-408.
Verplanken, B. (1991). Persuasive communication of risk information: A test of cue versus message
processing effects in a field experiment. Personality and Social Psychology Bulletin, 17, 188-193.
Vinokur, A., & Burnstein, E. (1974). The effects of partially shared persuasive arguments on groupinduced shifts: A group problem solving approach. Journal of Personality and Social Psychology, 29, 305-315.
Visser, P.S. & Krosnick, J. A. (1996). Age and susceptibility to attitude change: A new approach to an
old question. Unpublished manuscript. Ohio State University, Columbus, OH.
Watts, W. A. (1967). Relative persistence of opinion change induced by active compared to passive
participation. Journal of Personality and Social Psychology, 5, 4-15.
Watts, W. A., & Holt, L. E. (1979). Persistence of opinion change induced under conditions of
forewarning and distraction. Journal of Personality and Social Psychology, 37, 778-789.
Webster, D. M., & Kruglanski, A. W. (1995). Individual differences in need for cognitive closure.
Journal of Personality and Social Psychology, 67, 1049-1062.
Wegner, D. M. (1994). Ironic processes of mental control. Psychological Review, 101, 34-52.
Wegener, D. T., Downing, J., Krosnick, J. A., & Petty, R. E. (1995). Measures and manipulations of
strength-related properties of attitudes: Current practice and future directions. In R. E. Petty and J. A. Krosnick
(Eds.), Attitude strength: Antecedants and consequences (pp. 455-487). Mahwah, NJ: Erlbaum.
Wegener, D. T., & Petty, R. E. (1994). Mood-management across affective states: The hedonic
contingency hypothesis. Journal of Personality and Social Psychology, 66, 1034-1048.
Wegener, D. T., & Petty, R. E. (1995). Flexible correction processes in social judgment: The role of
naive theories in corrections for perceived bias. Journal of Personality and Social Psychology, 68, 36-51.
Wegener, D. T., & Petty, R. E. (1996). Effects of mood on persuasion processes: Enhancing, reducing,
and biasing scrutiny of attitude-relevant information. In L. L. Martin, & A. Tesser (Eds.). Striving and feeling:
Interactions between goals and affect. Mahwah, NJ: Erlbaum.
Wegener, D. T. & Petty, R. E. (1997). The flexible correction model: The role of naive theories in bias
correction. To appear in M. P. Zanna (Ed.), Advances in experimental social psychology.
Wegener, D. T., Petty, R. E., & Klein, D. J. (1994). Effects of mood on high elaboration attitude change:
The mediating role of likelihood judgments. European Journal of Social Psychology, 24, 25-43.
Wegener, D. T., Petty, R. E., & Smith, S. M. (1995). Positive mood can increase or decrease message
scrutiny: The hedonic contingency view of mood and message processing. Journal of Personality and Social
Psychology, 69, 5-15.
Weitzenhoffer, A. M., & Weitzenhoffer, G. B. (1958). Sex, transference, and susceptibility to hypnosis.
American Journal of Clinical Hypnosis, 1, 15-24.
Wells, G. L., & Petty, R. E. (1980). The effects of overt head movement on persuasion: Compatibility
and incompatibility of responses. Basic and Applied Social Psychology, 1, 219-230.
White, P. H., & S. G., Harkins (1994). Race of source effects in the Elaboration Likelihood Model.
Journal of Personality and Social Psychology, 67, 790-807.
Wilson, T. D., & Brekke, N. (1994). Mental contamination and mental correction: Unwanted influences
on judgments and evaluations. Psychological Bulletin, 116, 117-142.
Wilson, T. D., Dunn, D. S., Kraft, D., & Lisle, D. J. (1989). Introspection, attitude change, and attitudebehavior consistency: The disrupting effects of explaining why we feel the way we do. Advances in
Experimental Social Psychology, 22, 287-343.
Wilson, T. D., Kraft, D., & Dunn, D. S. (1989). The disruptive effects of explaining attitudes: The
moderating effect of knowledge about the attitude object. Journal of Experimental Social Psychology, 25,
379-400.
Witkin, H. A., Goodenough, D. R., & Oltman, P.K. (1979). Psychological differentiation: Current status.
Journal of Personality and Social Psychology, 37, 1127-1145.
Wood, W. (1982). Retrieval of attitude-relevant information from memory: Effects on susceptibility to
persuasion and on intrinsic motivation. Journal of Personality and Social Psychology, 42, 798-910.
Wood, W., & Kallgren, C. A. (1988). Communicator attributes and persuasion: Recipients access to
75
attitude-relevant information in memory. Personality and Social Psychology Bulletin, 14, 172-182.
Wood, W., Kallgren, C. A., & Preisler, R. M. (1985). Access to attitude-relevant information in memory
as a determinant of persuasion: The role of message attributes. Journal of Experimental Social Psychology, 21,
73-85.
Wood, W., Lundgren, S., Ouellette, J. A., Busceme, S., & Blackstone, T. (1994). Minority influence: A
meta-analytic review of social influence processes. Psychological Bulletin, 115, 323-345.
Wood, W., Rhodes, N., & Biek, M. (1995). Working knowledge and attitude strength: An information
processing analysis. In R. E. Petty, & J. A. Krosnick (Ed.), Attitude strength: Antecedents and consequences (pp.
283-313). Mahwah, NJ: Erlbaum.
Worth, L. T., & Mackie, D. M. (1987). Cognitive mediation of positive affect in persuasion. Social
Cognition, 5, 76-94.
Wright, P. L. (1981). Cognitive responses to mass media advocacy. In R. E. Petty, T. M. Ostrom, & T.
C. Brock (Eds.), Cognitive responses in persuasion. Hillsdale, NJ: Erlbaum.
Wu, C., & Shaffer, D. R. (1987). Susceptibility to persuasive appeals as a function of source credibility
and prior experience with the attitude object. Journal of Personality and Social Psychology, 52, 677-688.
Wyer, R. S., Jr. (1973). Category ratings as "subjective expected values": Implications for attitude
formation and change. Psychological Review, 80, 446-467.
Wyer, R. S., Jr. (1974). Cognitive organization and change: An information-processing approach.
Hillsdale, NJ: Erlbaum.
Wyer, R. S., Jr. (1991). The construction and use of thought systems: Some theoretical ambiguities. In R.
S. Wyer, & T. K. Srull (Ed.), Advances in Social Cognition (Vol. 4, pp. 203-214). Hillsdale, NJ: Erlbaum.
Yalch, R. F., & Elmore-Yalch, R. (1984). The effect of numbers on the route to persuasion. Journal of
Consumer Research, 11, 522-527.
Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology
Monograph Supplements, 9, 1-27.
Zajonc, R. B. (this volume). Affect, mood, and emotion. In D. Gilbert, S. Fiske, & G. Lindzey (Eds.)
Handbook of social psychology.
Zajonc, R. B., & Markus, H. (1982). Affective and cognitive factors in preferences. Journal of Consumer
Research, 9, 123-131.
Zanna, M. P. (1993). Message receptivity: A new look at the old problem of open- versus closedmindedness. In A. A. Mitchell (Ed.), Advertising exposure, memory and choice. Hillsdale, NJ: Erlbaum.
Zanna, M. P., & Cooper, J. (1974). Dissonance and the pill: An attribution approach to studying the
arousal properties of dissonance. Journal of Personality and Social Psychology, 29, 703-709.
Zanna, M. P., Kiesler, C. A., & Pilkonis, P. A. (1970). Positive and negative attitudinal affect established
by classical conditioning. Journal of Personality and Social Psychology, 14, 321-328.
Zanna, M. P., & Rempel, J. K. (1988). Attitudes: A new look at an old concept. In D. Bar-Tal, & A. W.
Kruglanski (Ed.), The social psychology of knowledge (pp. 315-334). Cambridge, England: Cambridge
University Press.
Zillmann, D. (1972). Rhetorical elicitation of agreement in persuasion. Journal of Personality and Social
Psychology, 21, 159-165.
Zimbardo, P. G. (1960). Involvement and communication discrepancy as determinants of opinion
conformity. Journal of Abnormal and Social Psychology, 60, 86-94.
Zimbardo, P. G., Ebbesen, E., & Maslach, C. (1977). Influencing attitudes and changing behavior.
Reading, MA: Addison-Wesley.
Zimbardo, P. G., Weisenberg, M., Firestone, I., & Levy, B. (1965). Communicator effectiveness in
producing public conformity and private attitude change. Journal of Personality, 33, 233-255.
Zuwerink, J. R., & Devine, P. G. (1996). Attitude importance and resistance to persuasion: It’s not just
the thought that counts. Journal of Personality and Social Psychology, 70, 931-944.
FOOTNOTES
76
1.. Discussions of the tradeoff hypothesis are not meant to imply that one could not construct
situations in which information relevant to a given heuristic or peripheral process is complex or
difficult to assess. In such a situation, one should find an impact of that peripheral process only if
people think enough to assess that information.
2.. McGuire (1995b) has recognized some limitations of this input-output matrix model. Among the
most notable limitations he mentions are that many of the model’s input factors have been shown to
interact rather than having main effects on the output or mediating factors, and that the model
“exaggerates the elaborateness with which audiences usually process persuasive communications (p.
235).” Nevertheless, the model can provides a useful way to organize a complex literature.
3.. Although these integration rules have generally been applied in attitude change settings to
integration of arguments in a persuasive messages, they can also be used to model the impact of
variables such as source credibility (Birnbaum, Wong, & Wong, 1976) and multiple sources
(Himmelfarb, 1972). Of course, the same difficulties regarding a priori specification of changes in the
weighting parameter also apply to these applications of the integration rules.
4.. If attributional processes were used to evaluate the central merits of an attitudinal position
rather than the characteristics or motivations of the source of the message, such processing would
have more in common with the high-elaboration processes discussed in the earlier section of this
chapter. For example, in some situations, consumers might make attributions about why advertisers
omit pieces of information that are noted by competitors (inferring that the product is inferior on that
feature; see Kardes, 1994).
5.. It is also possible that a variable is relevant to the central merits of the attitude object (e.g., the
attractiveness of the model in the current example) but under certain circumstances influences
persuasion through its operation as a simple cue -- as described later in the text. For example, if a
person receiving a communication about the potential model is distracted (and thus is unable to
extensively think about the merits of the person), then the attractiveness of the person might
influence the opinions of the message recipient through its impact as a simple cue (even though the
attractiveness of the person would have been considered more carefully as a central merit if the
message recipient had been able to scrutinize the information).
6.. Use of the word ?bias” is not meant to imply a necessary inaccuracy or incorrectness. Rather, the
term ?bias” is used to denote a situation in which one of a number of equally plausible interpretations
is consistently chosen based on the presence of absence of some other variable.
7.. Some researchers delineated a variety of types and sources of power. For example, French and
Raven (1959) compared reward power and coercive power (stemming from the ability to deliver
rewards and punishments, respectively), legitimate power (stemming from societal norms), referent
power (stemming from identification with the source), expert power (stemming from perceptions of the
source’s knowledge), and informational power (stemming from the strength of the arguments given by
the source). The first five types of power related to perceptions of the source him- or herself, but the
last form of power related purely to the information conveyed by the source. It has been more common
to discuss the effects of norms, identification with sources, expertise of sources, and information
presented by sources separate from the label of power (with discussions of power being limited
primarily to the ability to regulate rewards and punishments (i.e., reward and coercive power; see
Kelman, 1958). Our discussion in this section is limited to power with regard to control over positive
and negative outcomes, with discussions of the other characteristics of sources occurring under the
respective labels (e.g., source expertise).
77
8.. Smith and Shaffer (1995) actually showed an effect of both speech rate and argument quality
under low to moderate relevance, but only an effect of argument quality when relevance was high.
This is consistent with the notion of a continuum of elaboration in which the impact of peripheral cue
processes is maximal at the low-elaboration end of the continuum and decreases across the continuum
(until the high-elaboration end is reached, when the impact of peripheral processes is minimal; see
Petty & Cacioppo, 1986a; Petty et al., 1987).
9.. It is also possible that these effects are due to differences in perceived knowledge or expertise of the
in-group versus out-group source. The two topics used in the study were not rated by participants as
differentially important, though participants rated the in-group source as more qualified to speak to
the oil drilling issue than the out-group source (but not differentially qualified for the acid rain issue).
10.. In fact, in the literature on fear appeals, a number of operationalizations are used. In some, the
high fear message may include a greater number of negative consequences, or negative consequences
of greater severity than in the low fear message, or the same consequences may be implied but are
depicted more vividly, or are repeated more times in the high than low fear message. In still other
studies, the same message is given, but recipient reactions are assessed to determine fear. Or,
combinations of these features may be used to create high and low fear messages. Because of this
complexity and confounding, some fear studies are open to simple alternative interpretations (e.g., the
high fear message was more persuasive because it included more or better arguments).
11.. Some have argued that fear has opposite effects on some of the underlying processes of
persuasion (e.g., reducing reception but enhancing yielding). If so, then an inverted-U relationship
might be expected between fear and persuasion (e.g., Janis, 1967; McGuire, 1968). This has not
generally been observed (Boster & Mongeau, 1984), but perhaps the high levels of fear needed to
obtain it have not been present in the available research.
78